HRF Datasets Research Articles

Improving the Curvelet Saliency and Deep Convolutional Neural Networks for Diabetic Retinopathy Classification in Fundus Images

Retinal vessel images give a wide range of the abnormal pixels of patients. Therefore, classifying the diseases depending on fundus images is a popular approach. This paper proposes a new method to classify diabetic retinopathy in retinal blood vessel images based on curvelet saliency for segmentation. Our approach includes three periods: pre-processing of the quality of input images, calculating the saliency map based on curvelet coefficients, and classifying VGG16. To evaluate the results of the proposed method STARE and HRF datasets are used for testing with the Jaccard Index. The accuracy of the proposed method is about 98.42% and 97.96% with STARE and HRF datasets respectively.

Artery vein classification in fundus images using serially connected U-Nets

Background and objective: Retinal vessels provide valuable information when diagnosing or monitoring various diseases affecting the retina and disorders affecting the cardiovascular or central nervous systems. Automated retinal vessel segmentation can assist clinicians and researchers when interpreting retinal images. As there are differences in both the structure and function of retinal arteries and veins, separating these two vessel types is essential. As manual segmentation of retinal images is impractical, an accurate automated method is required.Methods: In this paper, we propose a convolutional neural network based on serially connected U-nets that simultaneously segment the retinal vessels and classify them as arteries or veins. Detailed ablation experiments are performed to understand how the major components contribute to the overall system’s performance. The proposed method is trained and tested on the public DRIVE and HRF datasets and a proprietary dataset.Results: The proposed convolutional neural network achieves an F1 score of 0.829 for vessel segmentation on the DRIVE dataset and an F1 score of 0.814 on the HRF dataset, consistent with the state-of-the-art methods on the former and outperforming the state-of-the-art on the latter. On the task of classifying the vessels into arteries and veins, the method achieves an F1 score of 0.952 for the DRIVE dataset exceeding the state-of-the-art performance. On the HRF dataset, the method achieves an F1 score of 0.966, which is consistent with the state-of-the-art.Conclusions: The proposed method demonstrates competitive performance on both vessel segmentation and artery vein classification compared with state-of-the-art methods. The method outperforms human experts on the DRIVE dataset when classifying retinal images into arteries, veins, and background simultaneously. The method segments the vasculature on the proprietary dataset and classifies the retinal vessels accurately, even on challenging pathological images. The ablation experiments which utilize repeated runs for each configuration provide statistical evidence for the appropriateness of the proposed solution. Connecting several simple U-nets significantly improved artery vein classification performance. The proposed way of serially connecting base networks is not limited to the proposed base network or segmenting the retinal vessels and could be applied to other tasks.

An Algorithm for Shrinking Blood Receptacles using Retinal Internal Pictures for Clinical Characteristics Measurement

The manual technique that might use for shrinking vessels blood in the retinal fundus images has significant limitations, such as the high rate of time consumption and the possibility of human error, precisely appear with the sophisticated structure of the blood receptacle and a hung amount of the retinal fundus photograph that needs to be anatomic. Moreover, the automatic proposed algorithm that will utilize shrinking and explore helpful clinical characteristics from retinal fundus photographs in order to lead the eye caregiver to early diagnosis for various retinal disorders and therapy evaluations. A precise, quick, and fully-automatic algorithm for shrinking blood receptacles and clinical characteristics measuring technique for internal retinal pictures is suggested in order to increase the diagnostic accuracy and reduce the ophthalmologist's burden. The proposed algorithm's main pipeline consists of two fundamental stages: picture shrinkage and medical feature elicitation. Many exhaustive practices were conducted to evaluate the efficacy of the sophisticated fully-automated shrinkage system in figuring out retinal blood receptacles using the DRIVE and HRF datasets of exceedingly demanding fundus images. Initially, the accuracy of the created algorithm was tested based on its ability to accurately recognize the retinal structure of blood receptacles. In these attempts, several quantitative performance measures precisely five were computed to validate the efficacy of the exact algorithm, including accuracy (Acc.), sensitivity (Sen.), specificity (Spe.), positive prediction value (PPV), and negative prediction value (NPV). When contrast with modern receptacles shrinking approaches on the DRIVE dataset, the produced results have enormously improved by obtaining accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of 98.78%, 98.32%, 97.23%, and 90. Based on five quantitative performance indicators, the HRF dataset led to the following results: 98.76%, 98.87%, 99.17%, 96.88%, and 100%.

Fundus Image Registration Technique Based on Local Feature of Retinal Vessels

Feature-based retinal fundus image registration (RIR) technique aligns fundus images according to geometrical transformations estimated between feature point correspondences. To ensure accurate registration, the feature points extracted must be from the retinal vessels and throughout the image. However, noises in the fundus image may resemble retinal vessels in local patches. Therefore, this paper introduces a feature extraction method based on a local feature of retinal vessels (CURVE) that incorporates retinal vessels and noises characteristics to accurately extract feature points on retinal vessels and throughout the fundus image. The CURVE performance is tested on CHASE, DRIVE, HRF and STARE datasets and compared with six feature extraction methods used in the existing feature-based RIR techniques. From the experiment, the feature extraction accuracy of CURVE (86.021%) significantly outperformed the existing feature extraction methods (p ≤ 0.001*). Then, CURVE is paired with a scale-invariant feature transform (SIFT) descriptor to test its registration capability on the fundus image registration (FIRE) dataset. Overall, CURVE-SIFT successfully registered 44.030% of the image pairs while the existing feature-based RIR techniques (GDB-ICP, Harris-PIIFD, Ghassabi’s-SIFT, H-M 16, H-M 17 and D-Saddle-HOG) only registered less than 27.612% of the image pairs. The one-way ANOVA analysis showed that CURVE-SIFT significantly outperformed GDB-ICP (p = 0.007*), Harris-PIIFD, Ghassabi’s-SIFT, H-M 16, H-M 17 and D-Saddle-HOG (p ≤ 0.001*).

DPN: detail-preserving network with high resolution representation for efficient segmentation of retinal vessels

Retinal vessels are important biomarkers for many ophthalmological and cardiovascular diseases. Hence, it is of great significance to develop automatic models for computer-aided diagnosis. Existing methods, such as U-Net follows the encoder-decoder pipeline, where detailed information is lost in the encoder in order to achieve a large field of view. Although spatial detailed information could be recovered partly in the decoder, while there is noise in the high-resolution feature maps of the encoder. And, we argue this encoder-decoder architecture is inefficient for vessel segmentation. In this paper, we present the detail-preserving network (DPN), which avoids the encoder-decoder pipeline. To preserve detailed information and learn structural information simultaneously, we designed the detail-preserving block (DP-Block). Further, we stacked eight DP-Blocks together to form the DPN. More importantly, there are no down-sampling operations among these blocks. Therefore, the DPN could maintain a high/full resolution during processing, avoiding the loss of detailed information. To illustrate the effectiveness of DPN, we conducted experiments over three public datasets. Experimental results show, compared to state-of-the-art methods, DPN shows competitive/better performance in terms of segmentation accuracy, segmentation speed, and model size. Specifically, (1) Our method achieves comparable segmentation performance on the DRIVE, CHASE_DB1, and HRF datasets. (2) The segmentation speed of DPN is over 20-160 $$\times$$ faster than other methods on the DRIVE dataset. (3) The number of parameters of DPN is around 120k, far less than all comparison methods.

BSEResU-Net: An attention-based before-activation residual U-Net for retinal vessel segmentation

Background and objectivesRetinal vessels are a major feature used for the physician to diagnose many retinal diseases, such as cardiovascular disease and Glaucoma. Therefore, the designing of an auto-segmentation algorithm for retinal vessel draw great attention in medical field. Recently, deep learning methods, especially convolutional neural networks (CNNs) show extraordinary potential for the task of vessel segmentation. However, most of the deep learning methods only take advantage of the shallow networks with a traditional cross-entropy objective, which becomes the main obstacle to further improve the performance on a task that is imbalanced. We therefore propose a new type of residual U-Net called Before-activation Squeeze-and-Excitation ResU-Net (BSEResu-Net) to tackle the aforementioned issues. MethodsOur BSEResU-Net can be viewed as an encoder/decoder framework that constructed by Before-activation Squeeze-and-Excitation blocks (BSE Blocks). In comparison to the current existing CNN structures, we utilize a new type of residual block structure, namely BSE block, in which the attention mechanism is combined with skip connection to boost the performance. What's more, the network could consistently gain accuracy from the increasing depth as we incorporate more residual blocks, attributing to the dropblock mechanism used in BSE blocks. A joint loss function which is based on the dice and cross-entropy loss functions is also introduced to achieve more balanced segmentation between the vessel and non-vessel pixels. ResultsThe proposed BSEResU-Net is evaluated on the publicly available DRIVE, STARE and HRF datasets. It achieves the F1-score of 0.8324, 0.8368 and 0.8237 on DRIVE, STARE and HRF dataset, respectively. Experimental results show that the proposed BSEResU-Net outperforms current state-of-the-art algorithms. ConclusionsThe proposed algorithm utilizes a new type of residual blocks called BSE residual blocks for vessel segmentation. Together with a joint loss function, it shows outstanding performance both on low and high-resolution fundus images.

Reliable IoT-based Health-care System for Diabetic Retinopathy Diagnosis to defend the Vision of Patients

PurposeThe purpose of this paper is to design an Internet-of-Things (IoT) architecture-based Diabetic Retinopathy Detection Scheme (DRDS) proposed for identifying Type-I or Type-II diabetes and to specifically advise the Type-II diabetic patients about the possibility of vision loss.Design/methodology/approachThe proposed DRDS includes the benefits of automatic calculation of clip limit parameters and sub-window for making the detection process completely adaptive. It uses the advantages of extended 5 × 5 Sobels operator for estimating the maximum edges determined through the convolution of 24 pixels with eight templates to achieve 24 outputs corresponding to individual pixels for finding the maximum magnitude. It enhances the probability of connecting pixels in the vascular map with its closely located neighbourhood points in the fundus images. Then, the spatial information and kernel of the neighbourhood pixels are integrated through the Robust Semi-supervised Kernelized Fuzzy Local information C-Means Clustering (RSKFL-CMC) method to attain significant clustering process.FindingsThe results of the proposed DRDS architecture confirm the predominance in terms of accuracy, specificity and sensitivity. The proposed DRDS technique facilitates superior performance at an average of 99.64% accuracy, 76.84% sensitivity and 99.93% specificity.Research limitations/implicationsDRDS is proposed as a comfortable, pain-free and harmless diagnosis system using the merits of Dexcom G4 Plantinum sensors for estimating blood glucose level in diabetic patients. It uses the merits of RSKFL-CMC method to estimate the spatial information and kernel of the neighborhood pixels for attaining significant clustering process.Practical implicationsThe IoT architecture comprises of the application layer that inherits the DR application enabled Graphical User Interface (GUI) which is combined for processing of fundus images by using MATLAB applications. This layer aids the patients in storing the capture fundus images in the database for future diagnosis.Social implicationsThis proposed DRDS method plays a vital role in the detection of DR and categorization based on the intensity of disease into severe, moderate and mild grades. The proposed DRDS is responsible for preventing vision loss of diabetic Type-II patients by accurate and potential detection achieved through the utilization of IoT architecture.Originality/valueThe performance of the proposed scheme with the benchmarked approaches of the literature is implemented using MATLAB R2010a. The complete evaluations of the proposed scheme are conducted using HRF, REVIEW, STARE and DRIVE data sets with subjective quantification provided by the experts for the purpose of potential retinal blood vessel segmentation.

A novel approach to glaucoma screening using optic nerve head through image fusion and fractal geometry

The Glaucoma is a typical eye issue that causes vision loss. It leads to visual impairment if it is untreated on time. Normally, vision loss is slow and not perceptible. Regular and systematic eye assessments are suggested for persons from middle age to prevent further vision loss. The proposed system introduced a new technique in the field of ophthalmology to diagnose glaucoma in an effective way using Image Fusion and Fractal Geometry techniques. The optic cup and disc are extracted from fundus images using K-means and Thresholding techniques. Optic cups generated by the techniques are combined to obtain the better cup region using the image Fusion Technique to improve the glaucoma screening process. The same process is applied on optic discs to obtain the fused disc area. The Box counting fractal dimension estimation technique from Fractal Geometry is applied on fused areas to classify the image as either healthy or glaucoma. Results of these two techniques are evaluated on a publically available HRF dataset and obtained the accuracy of 97%.

DRNet: Segmentation and localization of optic disc and Fovea from diabetic retinopathy image

Background and objectiveIn modern ophthalmology, automated Computer-aided Screening Tools (CSTs) are crucial non-intrusive diagnosis methods, where an accurate segmentation of Optic Disc (OD) and localization of OD and Fovea centers are substantial integral parts. However, designing such an automated tool remains challenging due to small dataset sizes, inconsistency in spatial, texture, and shape information of the OD and Fovea, and the presence of different artifacts. MethodsThis article proposes an end-to-end encoder-decoder network, named DRNet, for the segmentation and localization of OD and Fovea centers. In our DRNet, we propose a skip connection, named residual skip connection, for compensating the spatial information lost due to pooling in the encoder. Unlike the earlier skip connection in the UNet, the proposed skip connection does not directly concatenate low-level feature maps from the encoder's beginning layers with the corresponding same scale decoder. We validate DRNet using different publicly available datasets, such as IDRiD, RIMONE, DRISHTI-GS, and DRIVE for OD segmentation; IDRiD and HRF for OD center localization; and IDRiD for Fovea center localization. ResultsThe proposed DRNet, for OD segmentation, achieves mean Intersection over Union (mIoU) of 0.845, 0.901, 0.933, and 0.920 for IDRiD, RIMONE, DRISHTI-GS, and DRIVE, respectively. Our OD segmentation result, in terms of mIoU, outperforms the state-of-the-art results for IDRiD and DRIVE datasets, whereas it outperforms state-of-the-art results concerning mean sensitivity for RIMONE and DRISHTI-GS datasets. The DRNet localizes the OD center with mean Euclidean Distance (mED) of 20.23 and 13.34 pixels, respectively, for IDRiD and HRF datasets; it outperforms the state-of-the-art by 4.62 pixels for IDRiD dataset. The DRNet also successfully localizes the Fovea center with mED of 41.87 pixels for the IDRiD dataset, outperforming the state-of-the-art by 1.59 pixels for the same dataset. ConclusionAs the proposed DRNet exhibits excellent performance even with limited training data and without intermediate intervention, it can be employed to design a better-CST system to screen retinal images. Our source codes, trained models, and ground-truth heatmaps for OD and Fovea center localization will be made publicly available upon publication at GitHub.11https://github.com/kamruleee51/DRNet_Segmentation_Localization_OD_Fovea.

Retinal Image Analysis for Diabetes-Based Eye Disease Detection Using Deep Learning

Diabetic patients are at the risk of developing different eye diseases i.e., diabetic retinopathy (DR), diabetic macular edema (DME) and glaucoma. DR is an eye disease that harms the retina and DME is developed by the accumulation of fluid in the macula, while glaucoma damages the optic disk and causes vision loss in advanced stages. However, due to slow progression, the disease shows few signs in early stages, hence making disease detection a difficult task. Therefore, a fully automated system is required to support the detection and screening process at early stages. In this paper, an automated disease localization and segmentation approach based on Fast Region-based Convolutional Neural Network (FRCNN) algorithm with fuzzy k-means (FKM) clustering is presented. The FRCNN is an object detection approach that requires the bounding-box annotations to work; however, datasets do not provide them, therefore, we have generated these annotations through ground-truths. Afterward, FRCNN is trained over the annotated images for localization that are then segmented-out through FKM clustering. The segmented regions are then compared against the ground-truths through intersection-over-union operations. For performance evaluation, we used the Diaretdb1, MESSIDOR, ORIGA, DR-HAGIS, and HRF datasets. A rigorous comparison against the latest methods confirms the efficacy of the approach in terms of both disease detection and segmentation.

Automatic Diagnosis for Glaucoma Fusing Multi-Feature and Multi-Classifier

Glaucoma is a chronic and irreversible eye disease leading to blindness, and early detection is particularly important for its diagnosis and treatment. To improve the performance of automatic glaucoma diagnosis, a method based on multi-feature and multi-classifier is proposed. Firstly, an average histogram is obtained for each channel and ophthalmic condition, and 6 features are extracted from the average histogram with the average count of pixels and their maximum intensity value. Secondly, the optimal features combination is screened for each classifier with 10-fold cross-validation. Finally, the three optimal classifiers and their optimal features combination are fused with the principle of democratic voting. With the 10-fold cross-validation algorithm, the fusion model was evaluated on Local and HRF dataset, that achieved accuracy of 91.8% and 93.3%, sensitivity of 86.9% and 86.7%, specificity of 96.7% and 100%, AUC of 0.953 and 0.978, time cost of 1.0 s and 3.9 s per image, respectively. Simulation results show that the proposed method is of high accuracy and generality. It can effectively classify the retinal fundus images and provide technical support for the clinical diagnosis of retinal diseases.

Decision Support System for Diagnosing Diabetic Retinopathy from Color Fundus Images

Abstract Diabetic Retinopathy (DR) is a medical condition, also known as diabetic eye disease, which is vision-threatening damage to the retina of the eye caused by diabetes. As the technology advances, researchers are becoming more interested in intelligent medical diagnosis systems to assist screening of DR in earlier stages. In this study, variety of state-of-the-art procedures are used to extract the anatomic segments and lesions from the color fundus images. In addition, an automated system is proposed for the detection of anatomic segments and lesions by grading approach to help clinical diagnosis of the DR analysis. Four publicly available databases of color fundus images and various appropriate measurement techniques are used to compare quantitatively the performance of the proposed system. The experiments conducted on DIARETDB0, DIARETDB1, STARE, and HRF data sets have proved that accuracy, sensitivity, and specificity of the proposed system are comparable or superior to state-of-the-art methods.

Retinal blood vessel segmentation from diabetic retinopathy images using tandem PCNN model and deep learning based SVM

Diabetic Retinopathy (DR) occurs due to Type-II diabetes. It causes damages to the retinal blood vessels and reason for visual impairment. The predicted center is around the probability of variation in the estimation of retinal veins, and the crisp enrolls vessel development inside the retina. To witness the changes segmentation of retinal blood vessels has to be made. A framework to upgrade the quality of the segmentation results over morbid retinal images is proposed. This framework utilizes Contrast Limited Adaptive Histogram Equalization (CLAHE) for eliminating the background from the source image and enhances the foreground blood vessel pixels, Tandem Pulse Coupled Neural Network (TPCNN) model is endorsed for automatic feature vectors generation, and Deep Learning Based Support Vector Machine (DLBSVM) is proposed for classification and extraction of blood vessels. The DLBSVM parameters are fine-tuned via Firefly algorithm. The STARE, DRIVE, HRF, REVIEW, and DRIONS fundus image datasets are deliberated to assess the recommended techniques. The results render that the proposed technologies improve the segmentation with 80.61% Sensitivity, 99.54% Specificity, and 99.49% Accuracy.

Artery–vein segmentation in fundus images using a fully convolutional network

Epidemiological studies demonstrate that dimensions of retinal vessels change with ocular diseases, coronary heart disease and stroke. Different metrics have been described to quantify these changes in fundus images, with arteriolar and venular calibers among the most widely used. The analysis often includes a manual procedure during which a trained grader differentiates between arterioles and venules. This step can be time-consuming and can introduce variability, especially when large volumes of images need to be analyzed. In light of the recent successes of fully convolutional networks (FCNs) applied to biomedical image segmentation, we assess its potential in the context of retinal artery–vein (A/V) discrimination. To the best of our knowledge, a deep learning (DL) architecture for simultaneous vessel extraction and A/V discrimination has not been previously employed. With the aim of improving the automation of vessel analysis, a novel application of the U-Net semantic segmentation architecture (based on FCNs) on the discrimination of arteries and veins in fundus images is presented. By utilizing DL, results are obtained that exceed accuracies reported in the literature. Our model was trained and tested on the public DRIVE and HRF datasets. For DRIVE, measuring performance on vessels wider than two pixels, the FCN achieved accuracies of 94.42% and 94.11% on arteries and veins, respectively. This represents a decrease in error of 25% over the previous state of the art reported by Xu et al. (2017). Additionally, we introduce the HRF A/V ground truth, on which our model achieves 96.98% accuracy on all discovered centerline pixels. HRF A/V ground truth validated by an ophthalmologist, predicted A/V annotations and evaluation code are available at https://github.com/rubenhx/av-segmentation.

PoreNet: CNN-based Pore Descriptor for High-resolution Fingerprint Recognition

With the development of high-resolution fingerprint scanners, high-resolution fingerprint-based biometric recognition has received increasing attention in recent years. This paper presents a pore feature-based approach for biometric recognition. Our approach employs a convolutional neural network (CNN) model, DeepResPore, to detect pores in the input fingerprint image. Thereafter, a CNN-based descriptor is computed for a patch around each detected pore. Specifically, we have designed a residual learning-based CNN, referred to as PoreNet that learns distinctive feature representation from pore patches. For verification, the match score is generated by comparing pore descriptors obtained from a pair of fingerprint images in bi-directional manner using the Euclidean distance. The proposed approach for high-resolution fingerprint recognition achieves 2.91% and 0.57% equal error rates (EERs) on partial (DBI) and complete (DBII) fingerprints of the benchmark PolyU HRF dataset. Most importantly, it achieves lower FMR1000 and FMR10000 values than the current state-of-the-art approach on both the datasets.

Pore detection in high-resolution fingerprint images using deep residual network

This letter presents a residual learning-based convolutional neural network, referred to as DeepResPore, for detection of pores in high-resolution fingerprint images. Specifically, the proposed DeepResPore model generates a pore intensity map from the input fingerprint image. Subsequently, the local maxima filter is operated on the pore intensity map to identify the pore coordinates. The results of our experiments indicate that the proposed approach is effective in extracting pores with a true detection rate of 94:49% on Test set I and 93:78% on Test set II of the publicly available PolyU HRF dataset. Most importantly, the proposed approach achieves state-of-the-art performance on both test sets.

Fundus photography quality assessment based on topological extinction values

The quality of retina images is important for the diagnosis and treatment of eye diseases. Blur, distortion, low contrast, among other artifacts, inhibit the viewing of regions of interest. This work proposes a global descriptor based on two extinction values (number of descendants and topological height) for direct analysis of fundus photographs and automatic classification as good or poor quality. The method does not require any special filtering, image segmentation or internal retina structure location. It is not substantially affected by rotating, resizing or data set type, and is computed in quadratic time. The algorithm achieved an area under ROC curve of 96.13% for UNICAMP, 88.59% for UNIFESP DR2, 99.27% for DRIMDB, and 92.07% for HRF data sets.

Fundus photography quality assessment based on topological extinction values

The quality of retina images is important for the diagnosis and treatment of eye diseases. Blur, distortion, low contrast, among other artifacts, inhibit the viewing of regions of interest. This work proposes a global descriptor based on two extinction values (number of descendants and topological height) for direct analysis of fundus photographs and automatic classification as "good" or "poor" quality. The method does not require any special filtering, image segmentation or internal retina structure location. It is not substantially affected by rotating, resizing or data set type, and is computed in quadratic time. The algorithm achieved an area under ROC curve of 96.13% for UNICAMP, 88.59% for UNIFESP DR2, 99.27% for DRIMDB, and 92.07% for HRF data sets.

A Coarse-to-Fine Fully Convolutional Neural Network for Fundus Vessel Segmentation

Fundus vessel analysis is a significant tool for evaluating the development of retinal diseases such as diabetic retinopathy and hypertension in clinical practice. Hence, automatic fundus vessel segmentation is essential and valuable for medical diagnosis in ophthalmopathy and will allow identification and extraction of relevant symmetric and asymmetric patterns. Further, due to the uniqueness of fundus vessel, it can be applied in the field of biometric identification. In this paper, we remold fundus vessel segmentation as a task of pixel-wise classification task, and propose a novel coarse-to-fine fully convolutional neural network (CF-FCN) to extract vessels from fundus images. Our CF-FCN is aimed at making full use of the original data information and making up for the coarse output of the neural network by harnessing the space relationship between pixels in fundus images. Accompanying with necessary pre-processing and post-processing operations, the efficacy and efficiency of our CF-FCN is corroborated through our experiments on DRIVE, STARE, HRF and CHASE DB1 datasets. It achieves sensitivity of 0.7941, specificity of 0.9870, accuracy of 0.9634 and Area Under Receiver Operating Characteristic Curve (AUC) of 0.9787 on DRIVE datasets, which surpasses the state-of-the-art approaches.

Learning audio and image representations with bio-inspired trainable feature extractors

Since when very young, we can quickly learn new concepts, and distinguish between different kinds of object or sound. If we see a single object or hear a particular sound, we are then able to recognize such sample or even different versions of it in other scenarios. As an example, if one sees a iron chair and associates the object to the general concept of “chairs”, he will be able to detect and recognize also wooden or wicker chairs. Similarly, when we hear the sound of a particular event, such as a scream, we are then able to recognize other kinds of scream that occur in different environments. We continuously learn representations of the real world, which we then use in order to understand new and changing environments.In the field of pattern recognition, traditional methods typically require a careful design of data representations (i.e. features), which involves considerable domain knowledge and effort by experts. Recently, approaches for automated learning of representations from training data were introduced and based on popular deep learning techniques and convolutional neural networks (CNN). Representation learning aims at avoiding engineering of hand-crafted features and providing automatically learned features suitable for the recognition tasks. In this work, we proposed novel trainable filters for representation learning in audio and image processing. The structure of these filters is not fixed in the implementation but rather configured directly from single prototype patterns of interest [4].In the context of audio processing, we focused on the problem of audio event detection and classification in noisy environments, also in cases where the signal to noise ratio (SNR) is null or negative. We released two data sets, namely the MIVIA audio events and the MIVIA road events data sets, and obtained baseline results (recognition rate of about 85%) with a real-time method for event detection based on the bag of features classification scheme [3, 2].We designed novel trainable feature extractors, which we call COPE (Combination of Peaks of Energy), that are able to detect specific constellations of energy peak points in time-frequency representations of input audio signals [8]. The particular constellation of energy peaks to be detected by a COPE feature extractor is determined in an automatic configuration process performed on a given prototype sound. The design of COPE feature extractors was inspired by some functions of the cochlea membrane and the inner hair cells in the inner auditory system, which convert the sound pressure waves into neural stimuli on the auditory nerve.We proposed a method that uses COPE feature extractors together with a classification system to perform real-time audio event detection and classification, also in cases where sounds have null and negative SNR. The performance results (recognition rate over 90%) that we obtained on several benchmarking data sets for audio events detection in different contexts are higher than state-of-the-art approaches.In the second part of the work, we introduced B-COSFIRE filters for detection of elongated and curvilinear patterns in images and apply them to the delineation of blood vessels in retinal images [1, 6]. The B-COSFIRE filters are trainable, that is their structure is automatically configured from prototype elongated patterns. The design of the B-COSFIRE filters is inspired by the functions of some neurons, called simple cells, in area V1 of the visual system, which fire when presented with line or contour stimuli. A B-COSFIRE filter achieves orientation selectivity by computing the weighted geometric mean of the output of a pool of Difference-of-Gaussians (DoG) filters, whose supports are aligned in a collinear manner. Rotation invariance is efficiently obtained by appropriate shiftings of the DoG filter responses.After configuring a large bank of B-COSFIRE filters selective for vessels (i.e. lines) and vessel-endings (i.e. line-endings) of various thickness (i.e. scale), we employed different techniques based on information theory and machine learning to select an optimal subset of B-COSFIRE filters for the vessel delineation task [5, 7]. We considered the selected filters as feature extractors to construct a pixel-wise feature vector which we used in combination with a classifier to classify the pixels in the testing image as vessel and non-vessel pixels. We carried out experiments on public benchmarking data sets (DRIVE, STARE, CHASE DB1 and HRF data sets) and the results that we achieved are higher than many existing methods.We studied the computational requirements of the proposed algorithms in order to evaluate their applicabilityin real-world applications and the fulfillment of real-time constraints given by the considered problems. The MATLAB implementation of the proposed algorithms are publicly released for research purposes.This work contributes to the development of algorithms for representation learning in audio and imageprocessing and promotes their use in higher-level pattern recognition systems.