Text Regions Research Articles

Robust Texture-Aware Local Adaptive Image Watermarking With Perceptual Guarantee

Watermarking involves embedding a watermark in an image and later extracting it to prove the image’s copyright. In most cases, a complete image contains both smooth and textured regions. As a rule of thumb, the visual quality of an image with a watermark embedded in its textured regions is better than that of the same image with a watermark in smooth regions. This paper, by taking advantage of the fact, proposes a texture-aware local adaptive watermarking algorithm to maximize the watermark’s robustness while maintaining its imperceptibility. To identify textured regions in an image, we introduce the texture value, an efficient and proper metric of the richness of image texture. It combines the texture correlation of the AC coefficients, the luminance masking of the DC coefficient, and the distribution of image texture. A watermark is embedded adaptively into multiple non-overlapping textured regions of an image under the specified SSIM condition. Its adaptiveness comes from a novel texture-aware adaptive parameter model derived by multivariate regression analysis. Correct extraction of watermarks from multiple textured regions can be done by the cooperation of embedding and extraction strategies, with the assistance of RS-based watermark coding model. They allow for greater robustness, faster extraction, and adjustable watermark capacity. The simulation experiments on 100 images demonstrate that our proposed algorithm outperforms state-of-the-art algorithms with respect to imperceptibility, robustness, and adaptability.

Advancements in OCR: A Deep Learning Algorithm for Enhanced Text Recognition

Optical Character Recognition (OCR) has significantly evolved with the rise of deep learning techniques. In this research paper, we present a novel and advanced OCR algorithm that leverages the power of deep learning for improved text recognition accuracy. Traditional OCR methods have faced limitations in handling complex layouts, noisy images, and diverse fonts, affecting overall performance. Our proposed algorithm addresses these challenges through the integration of deep neural networks, specifically convolutional and recurrent layers. The algorithm undergoes comprehensive training on large-scale datasets, enabling it to learn intricate patterns and features, resulting in robust recognition capabilities. Furthermore, we introduce an attention mechanism that enhances the model's ability to focus on critical text regions, enhancing accuracy and efficiency. Through extensive experiments and evaluations on benchmark datasets, we demonstrate the superiority of our deep learning-based OCR algorithm over conventional approaches. Our algorithm achieves state-of-the-art performance on various OCR tasks, including multilingual text recognition and document digitization. Additionally, we conduct an in-depth analysis of the algorithm's behaviour under various scenarios, such as low-resolution inputs and challenging environmental conditions. The findings from this research not only contribute to the field of OCR but also open avenues for applications in document analysis, text extraction, and content digitization in real-world scenarios. The integration of deep learning in OCR showcases its potential in revolutionising text recognition tasks, pushing the boundaries of accuracy and efficiency in this domain.

FMAP: Learning robust and accurate local feature matching with anchor points

Local feature matching involves the task of establishing the pixel-wise correspondences between a pair of images. As an integral component of plentiful computer vision applications (e.g., visual localization), this task has been successfully performed using Transformer-based methods. However, these methods typically extract numerous keypoints from sparse texture regions to construct a densely-connected graph neural network (GNN) for long-range feature aggregation, which inevitably triggers redundant message exchange and hampers the learning process. Furthermore, they employ transformer encoder layers that consider images as 1D sequences, leaving them incapable of extracting multiscale local structural information from the images, which is critical for establishing correspondence in image pairs with significantly scales shifts. In this study, we develop FMAP, an innovative detector-free approach that enables accurate local feature matching. For the first issue, FMAP employs an anchor points feature aggregation module (APAM) that captures representative keypoints and discards the extraneous keypoints to build a sparsified GNN for compact yet clean message exchange, with the key insight that the keypoints containing abundant visual information are distinguishable from their neighbors. For the second issue, FMAP proposes a global–local multiscale perception module (GMPM), which incorporates abundant multiscale local context information into global feature representation by employing multiple depth-wise convolutions with varying kernel sizes, thereby generating discriminative features that are robust to scale shifts. In addition, the depth-wise convolutions are utilized in the feed-forward network of the GMPM to further fuse the global context information and local feature representation. Extensive experiments on several standard benchmarks demonstrate that the proposed FMAP method significantly outperforms state-of-the-art methods. Compared to the cutting-edge methods MatchFormer, QuadTree, and TopicFM in relative pose estimation task, FMAP surpasses them by 2.27%, 0.58%, and 1.08% in terms of AUC@5°. Besides, FMAP noticeably outperforms the baseline LoFTR by (2.38%,1.89%,1.45%) in terms of AUC@(5°, 10°, 20°). Moreover, we integrate FMAP into an official visual localization framework and conduct a visual localization experiment, with the results showing that FMAP exceeds LoFTR by 2.3% in terms of AP.

Complex Text Detection Algorithm Based on Edge Attention Mechanism

In computer vision, automatically identifying and locating text in images or videos is an important task. Traditional text detection methods are not effective in detecting complex scenes in irregular rectangular areas such as text bending, spacing, and special shapes, which are mainly reflected in the fragmentation of text detection areas. In this paper, a text detection method based on edge attention mechanism is proposed to better adapt to complex scenes. The proposed method takes Encoder–Decoder as the core idea. First of all, an edge attention module is designed, including global attention and local attention. The global attention module is used to perceive the features of text regions and nontext regions, while the local attention module is used to learn the information of text boundaries. Then a multi-scale feature fusion process is designed, which can strengthen the edge information and key information of text regions. Finally, the model outputs probability maps and threshold maps, and generates high-precision binary maps of text regions. After experimental verification, the proposed method on the public data set significantly reduces the fragmentation of the detection area, improves the detection accuracy of the text area, and has better robustness for text detection scenes with unconventional rectangular areas.

Root‐zone enrichment of soil‐test biological activity and particulate organic carbon and nitrogen under conventional and conservation land management

AbstractBiologically active and intermediately labile fractions of soil organic matter are important features of conservation agricultural management systems. Depth distribution of soil‐test biological activity (STBA) and particulate organic carbon (C) and nitrogen (N) were characterized under cropland, grassland, and woodland across 25 research stations in North Carolina. Most (87%) of the 310 fields sampled were on Ultisols (udults or aquults). Variation in STBA in the surface 30‐cm of soil was controlled primarily by land use and to a lesser extent by soil texture and physiographic region. Baseline concentrations of STBA and particulate organic C and N at 30‐cm depth were relatively narrowly constrained compared with contents in the 0‐ to 10‐cm depth. Root‐zone enrichment (i.e., total minus baseline contents) of STBA (mean ± standard error; kg CO2–C ha−1 3 d−1) followed the order: conventional‐till cropland (194 ± 17) < no‐till cropland (274 ± 26) < woodland (421 ± 17) < grassland (537 ± 15) across locations. Proportion of root‐zone enrichment of soil organic C as particulate organic C was 43 ± 2%. Root‐zone enrichment of particulate organic C (Mg C ha−1) followed the order: conventional‐till cropland (3.8 ± 0.6) < no‐till cropland (8.7 ± 0.8) < grassland (14.1 ± 0.5) < woodland (16.3 ± 0.6) across locations. The C:N ratio of root‐zone‐enriched particulate organic matter was greater under woodland than under other land uses likely because of N input limitations. Conservation land uses increased root‐zone enrichment of these active and intermediately labile fractions independent of soil types and regions in the state.

Text line extraction strategy for palm leaf manuscripts

Text line segmentation is an important step in the historical document image analysis pipeline to supply useful information for recognition, keyword spotting and indexing. Many handcrafted-based and learning-based approaches have been developed to cope with text line extraction challenges. In this work we present a hybrid technique which combines a convolutional-based denoising task and heuristic Seam Carving framework. We propose the following changes to the original Seam Carving: (1) We applied adaptive slice to anticipate miss-extraction on a short text during medial seam computation. (2) We applied a triple smoothing to find the best local maxima of the smoothed horizontal projection profile which represents candidate medial seams. (3) We utilized five post-processing steps aimed at reconstructing a more precise medial seam. Three different palm leaf data sets: old Sundanese, old Balinese, and old Khmer, have been used to compare Convolutional Seam Carving (CSC) to several baseline methods, including the original Seam Carving. Experimental results show that the proposed method outperforms other current handcrafted-based baselines on all three palm leaf manuscript (PLM) data sets. On the old Sundanese data sets, CSC can produce a significant improvement of the performance rate compared to all other baseline approaches, and it also enhance the measurement on old Balinese and old Khmer datasets. In the ablation study, we discovered that a foreground extraction step is not only able to reduce noises and color degradation but also provide better separation of text region. Following that, an adaptive slice and triple smoothing approach contribute to solve the text length variation problem. Finally, post-processing steps were effective in connecting discontinuous medial seam. The code has been published in https://github.com/erickpaulus/text-line-segmentation.

Steganography for medical record image

Medical record images in EHR system are users’ privacy and an asset, and there is an urgent need to protect this data. Image steganography can offer a potential solution. A steganographic model for medical record images is therefore developed based on StegaStamp. In contrast to natural images, medical record images are document images, which can be very vulnerable to image cropping attacks. Therefore, we use text region segmentation and watermark region localization to combat the image cropping attack. The distortion network has been designed to take into account the distortion that can occur during the transmission of medical record images, making the model robust against communication induced distortions. In addition, based on StegaStamp, we innovatively introduced FISM as part of the loss function to reduce the ripple texture in the steganographic image. The experimental results show that the designed distortion network and the FISM loss function term can be well suited for the steganographic task of medical record images from the perspective of decoding accuracy and image quality.

A survey of text detection and recognition algorithms based on deep learning technology

Optical Character Recognition (OCR) poses a crucial challenge within the realm of computer vision research, as it plays a pivotal role in converting vast amounts of unstructured text data into structured formats to support diverse artificial intelligence applications. The OCR process encompasses two core components: text detection and text recognition. Text detection involves identifying and extracting text regions, achieved through either object detection or segmentation techniques, while text recognition focuses on accurately deciphering the content within these identified regions. In recent years, remarkable strides have been made in the domain of text recognition, primarily driven by deep learning-based models. These models eliminate the need for manual feature processing and excel in recognizing text even within complex scenes, surpassing the performance of traditional text recognition methods and subsequently emerging as the dominant approach. The objective of this paper is to present a comprehensive survey of both text detection and text recognition models. Firstly, we systematically categorize and provide an overview of existing off-the-shelf text detection methods. Subsequently, we conduct an in-depth investigation of six distinct text recognition models, taking into account their unique implementations. Additionally, we explore and analyze the principal datasets that currently prevail in the field of text detection and recognition. Furthermore, this research entails a meticulous performance comparison of various text detection algorithms on the CTW1500, TotalText, and ICDAR2015 datasets. Additionally, we evaluate and scrutinize the efficacy of mainstream text recognition algorithms on the IIIT-5K, SVT, ICDAR2013, SVT-P, CUTE80, and ICDAR2015 datasets. Finally, we conclude with a discussion on the future development and research trends concerning text detection and recognition, providing insights that can further drive progress in this crucial area.

Optimal Classification Model for Text Detection and Recognition in Video Frames

Currently, the identification of text from video frames and normal scene images has got amplified awareness amongst analysts owing to its diverse challenges and complexities. Owing to a lower resolution, composite backdrop, blurring effect, color, diverse fonts, alternate textual placement among panels of photos and videos, etc., text identification is becoming complicated. This paper suggests a novel method for identifying texts from video with five stages. Initially, “video-to-frame conversion”, is done during pre-processing. Further, text region verification is performed and keyframes are recognized using CNN. Then, improved candidate text block extraction is carried out using MSER. Subsequently, “DCT features, improved distance map features, and constant gradient-based features” are extracted. These characteristics subsequently provided “Long Short-Term Memory (LSTM)” for detection. Finally, OCR is done to recognize the texts in the image. Particularly, the Self-Improved Bald Eagle Search (SI-BESO) algorithm is used to adjust the LSTM weights. Finally, the superiority of the SI-BESO-based technique over many other techniques is demonstrated.

Feature fusion-based text information mining method for natural scenes

Abstract As a crucial medium of information dissemination, text holds a pivotal role in a multitude of applications. However, text detection in complex and unstructured environments presents significant challenges, such as the presence of cluttered backgrounds, variations in appearance, and uneven lighting conditions. To address this issue, this study proposes a text detection framework that leverages multistage edge detection and contextual information. This framework deviates from traditional approaches by incorporating four primary processing steps, including text visual saliency region detection to accentuate the text regions and diminish background interference, multistage edge detection to enhance the conventional stroke width transform results, a texture-based and connected components-based integration to accurately distinguish text from the background, and a context fusion step to recover missing text regions and improve the recall of text detection. The proposed method was evaluated on two widely used benchmark datasets, i.e., the international conference on document analysis and recognition (ICDAR) 2005 dataset and the ICDAR 2011 dataset, and the results indicate the advancedness of the method.

Digital Image Text Recognition Using Machine Learning Algorithms

Abstract: Text recognition is used in various fields like document analysis, picture labeling, and text analysis, etc. The process of recognizing text from image is very crucial.. Employing Maximally Stable Extremal Regions (MSER) and Optical Character Recognition (OCR) algorithms have drastically changed the accuracy and reliability of text detection from photos. Our proposed work suggests a model for text recognition that makes use of both the MSER and OCR algorithms' benefits for the purpose of accuracy improvement and reliability of the digital text extraction procedure. OCR follows cutting-edge techniques to distinguish and identify individual characters, serving as the essential building block for text recognition. However, these algorithms may struggle when dealing with complex backdrops, blurry photos, or text that is structured in an atypical way. We use MSER approach, which excels at recognizing text sections by finding maximally stable regions across various severities and scales, to solve these constraints. The suggested model employs a multi-stage methodology. The MSER algorithm is used for extracting the likely text spots from the input image first. To boost OCR performance, these zones are then fine-tuned using pre-processing techniques including noise reduction and picture enhancement. The OCR system next processes the cleaned-up sections, identifying each region's text using machine learning and pattern recognition methods. The text that is recognized is then further processed to increase the accuracy and refine these findings. Thus, compared to most other models, the text recognition model that is built utilizing the MSER and the CNN (OCR, a component of CNN) algorithm performs better

Text Recognition from Images using CNN and MSER Algorithms

Abstract: The ability to recognize text from images is a great importance in a range of applications, including document analysis, images captioning and augmented reality. The reliability and accuracy of text extraction from images have been completely transformed by text recognition models using Optical Character Recognition (OCR) and Maximally Stable Extremal Regions (MSER) algorithms. In our study, we propose a text recognition model that leverages the advantages of both OCR and MSER algorithms to enhance the reliability and accuracy of the text extraction process. OCR algorithm serve as the fundamental basis for text recognition by utilizing advanced methods to separate and identify individual characters. None the less these algorithms can encounter difficulties when confronted with intricated backgrounds, images of poor quality, or text arranged in irregular layouts. To address these limitations, we integrate the MSER algorithm, which excels in detecting text regions by identifying maximally stable regions across different scales and intensities. Our proposed model follows a multi-stage approach. First, the input image, the MSER method is used to extract probable text locations. These regions are then refined using pre-processing techniques, such as noise removal and image enhancement, to improve OCR performance. Next, the refined regions are passed through the OCR algorithm, which utilizes machine learning and pattern recognition techniques to recognize the text within each regions. The recognized text is subsequently post-processed to refine the results and improve overall accuracy. The text recognition model is implemented using CNN (OCR which is part of CNN) and the Maximally stable Extremal regions (MSER) algorithms

UTTSR: A Novel Non-Structured Text Table Recognition Model Powered by Deep Learning Technology

To prevent the compilation of documents, many table documents are formatted with non-editable and non-structured texts such as PDFs or images. Quickly recognizing the contents of tables is still a challenge due to factors such as irregular formats, uneven text quality, and complex and diverse table content. This article proposes the UTTSR table recognition model, which consists of four parts: text region detection, text line detection and recognition, and table sequence recognition. For table detection, the Cascade Faster RCNN with the ResNeXt105 network is implemented, using TPS (Thin Plate Spline) transformation and affine transformation to correct the image and to improve accuracy. For text line detection, DBNET is used with Do-Conv in FPN (Feature Pyramid Networks) to speed up training. Text lines are recognized using CRNN without the CTC module, enhancing recognition performance. Table sequence recognition is based on the transformer combined with post-processing algorithms that fuse table structure sequences and unit grid content. Experimental results show that the UTTSR model outperforms the compared methods. This upgraded model significantly improves the accuracy of the previous state-of-the-art F1 score on complex tables, reaching 97.8%.

Exploring Stroke-Level Modifications for Scene Text Editing

Scene text editing (STE) aims to replace text with the desired one while preserving background and styles of the original text. However, due to the complicated background textures and various text styles, existing methods fall short in generating clear and legible edited text images. In this study, we attribute the poor editing performance to two problems: 1) Implicit decoupling structure. Previous methods of editing the whole image have to learn different translation rules of background and text regions simultaneously. 2) Domain gap. Due to the lack of edited real scene text images, the network can only be well trained on synthetic pairs and performs poorly on real-world images. To handle the above problems, we propose a novel network by MOdifying Scene Text image at strokE Level (MOSTEL). Firstly, we generate stroke guidance maps to explicitly indicate regions to be edited. Different from the implicit one by directly modifying all the pixels at image level, such explicit instructions filter out the distractions from background and guide the network to focus on editing rules of text regions. Secondly, we propose a Semi-supervised Hybrid Learning to train the network with both labeled synthetic images and unpaired real scene text images. Thus, the STE model is adapted to real-world datasets distributions. Moreover, two new datasets (Tamper-Syn2k and Tamper-Scene) are proposed to fill the blank of public evaluation datasets. Extensive experiments demonstrate that our MOSTEL outperforms previous methods both qualitatively and quantitatively. Datasets and code will be available at https://github.com/qqqyd/MOSTEL.

DenseTextPVT: Pyramid Vision Transformer with Deep Multi-Scale Feature Refinement Network for Dense Text Detection.

Detecting dense text in scene images is a challenging task due to the high variability, complexity, and overlapping of text areas. To adequately distinguish text instances with high density in scenes, we propose an efficient approach called DenseTextPVT. We first generated high-resolution features at different levels to enable accurate dense text detection, which is essential for dense prediction tasks. Additionally, to enhance the feature representation, we designed the Deep Multi-scale Feature Refinement Network (DMFRN), which effectively detects texts of varying sizes, shapes, and fonts, including small-scale texts. DenseTextPVT, then, is inspired by Pixel Aggregation (PA) similarity vector algorithms to cluster text pixels into correct text kernels in the post-processing step. In this way, our proposed method enhances the precision of text detection and effectively reduces overlapping between text regions under dense adjacent text in natural images. The comprehensive experiments indicate the effectiveness of our method on the TotalText, CTW1500, and ICDAR-2015 benchmark datasets in comparison to existing methods.

Effective contact texture region aware pavement skid resistance prediction via convolutional neural network

AbstractThe surface texture of asphalt pavement has a significant effect on skid resistance performance. However, its contribution to the performance of skid resistance is non‐homogeneous and subjects to local validity. There are also a few deep learning models that take into account the effective contact texture region. This paper proposes a convolutional neural network model based on the effective contact texture region, containing macro‐ and micro‐scale awareness sub‐modules. In this study, the asphalt mixture with varying gradations was designed to accurately obtain the effective contact texture region. Then, the textures were disentangled into macro‐ and micro‐texture scales by applying the fast Fourier transform and fed into the model for training. Finally, the area of effective contact texture region was calculated, and the effective contact ratio parameter was then proposed using the triangulation algorithm. The results showed that the effective contact texture area of pavement varies by the asphalt mixture type. The effective contact ratio parameter exhibited a significant positive correlation (Pearson correlation coefficient is 0.901, R2= 0.8129) with skid resistance performance and was also influenced by key sieve aggregate content from 2.36 to 4.75 mm. The data of effective contact texture region following disentanglement significantly released the model performance (the relative error dropped to 1.81%). The model exhibited improved precision and performance, which can be utilized as an efficient, non‐contact alternative method for skid resistance analysis.

Perceiving Multiple Representations for scene text image super-resolution guided by text recognizer

Single image super-resolution (SISR) aims to recover clear high-resolution images from low-resolution images, which has made great progress with the development of deep learning these years. Scene text image super-resolution (STISR) is a subfield of SISR with the goal of increasing the resolution of a low-resolution text image and enhancing the readability of characters in the image. Despite significant improvements in recent approaches, STISR remains a challenging task due to the diversity of background, text appearances and layouts, etc. This paper presents a Perceiving Multiple Representations (PerMR) method for better super-resolution performances in scene text images. PerMR is a unified network that combines super-resolution with text recognition and exploits the recognizer’s feedback to facilitate super-resolution. Specifically, contextual information from the text decoder is extracted to provide sequence-specific guidance and enable the super-resolution model to pay more attention to the text region. Meanwhile, low-level and high-level visual features from the vision backbone of the recognition network are integrated to further improve visual quality. Additionally, we incorporate a frequency branch into the vanilla convolution unit, which efficiently enhances global and local feature representations. Experiments on the STISR benchmark dataset TextZoom validate that PerMR can not only generate more distinguishable images, but also outperforms the current state-of-the-art methods. PerMR boosts the average recognition accuracy by 5.9% using ASTER, 5.8% using MORAN and 10.6% using CRNN compared to the baseline model TSRN. PerMR outperforms the advanced method TPGSR-3 by 1.4% on ASTER, 0.1% on MORAN, 0.2% on CRNN and boosts TATT by 0.6% on ASTER and 1.1% on MORAN respectively. Furthermore, PerMR demonstrates good robustness and generalization when tackling low-quality text images in multiple scene text recognition datasets. The experiment results verify the capabilities of PerMR to boost text recognition performance.

Directed searching optimized texture based adaptive gamma correction (DSOTAGC) technique for medical image enhancement.

Because of complexity and low contrast in medical images, few enhancement techniques result unwanted artifacts and information loss by affecting the structure similarity and peak signal to noise ratio. To meet these challenges, a Directed searching optimized texture-based adaptive gamma correction technique is proposed in this article. This proposed technique utilizes the textured regions of the image and suppresses the effect of non-textured regions for eliminating the artifacts. An adaptive clipping threshold is used in the textured image to control the enhancement rate. For improving the contrast, the transfer function of the enhanced image is evaluated using the modified weighted probability density function and adaptive gamma parameter. To make the algorithm more adaptive, parameters like clipped threshold, gamma parameter, and textural threshold are to be optimized using directed searching optimization algorithm. For improving the information contents and noise suppression capability, the proposed technique incorporated a fitness function which is a combination of entropy and peak signal to noise ratio. Equal weightage has been given to each parameter in the fitness function for obtaining a balanced optimal result. Then, the performance of the proposed technique is evaluated in terms of visual quality, information contents, average mean brightness error, noise suppression, and structural similarity. Experimental results show the proposed technique results in better visual effects without information loss. It effectively suppresses the effect of artifacts and significantly improves the contrast by making edges clearer and textures richer over other algorithms.

Enhanced grasshopper optimization‐based selection of ultrasound and elastography features for breast lesion classification

AbstractBreast cancer (BC) is the second leading cause of death in women worldwide, caused by the uncontrolled proliferation of malignant cells in the breast. It is very tough to differentiate cancer from non‐cancer cells because of their shape and structure. Therefore, early diagnosis and classification of breast tumors as benign is a hot research topic. However, existing methods for early diagnosis have shortcomings like high computational cost and low classification accuracy. To trounce those limitations, this paper presented a computer‐aided diagnosis (CAD) system employing metaheuristic algorithms and machine learning classifiers. The proposed system uses fused features of ultrasound (US) and elastography images to enhance classification accuracy. Firstly, both US and elastography images are separately preprocessed using an enhanced wavelet thresholding method to diminish speckle noise. Secondly, a new segmentation algorithm is proposed to extract the lesion region from the preprocessed US image. Then the contour of the segmented lesion is mapped on an elastography image to obtain the corresponding region of interest (ROI), and texture and morphology features are separately extracted and concatenated. Subsequently, the enhanced grasshopper optimization algorithm (EGOA) is adopted to select salient features and reduce the dimension. Finally, based on the selected features, two machine learning classifiers, linear discriminant analysis (LDA) and support vector machine (SVM), are employed to classify the lesions into benign and malignant. The effectiveness of the proposed method is validated on 520 tumor images, including both benign and malignant samples. The suggested performance has been assessed by computing the confusion matrix and receiver operating characteristic curve. The developed method, EGOA‐SVM, outperforms the other classifiers by reaching a higher classification rate of 99.4%, specificity of 100%, the sensitivity of 99.2%, and Matthew's coefficient of 0.979. The introduced method can improve the radiologist's performance in categorizing breast lesions with a good classification rate. It could be a promising tool for clinical use.

Cork classification based on multi-scale faster-RCNN with machine vision

In traditional industrial production, the classification of cork mainly adopts manual operation. Due to round cork blocks with complex textures, manual sorting is time-consuming and often leads to inefficiency. It will lead to inefficiency and a high error rate. Therefore, an automatic cork sorting system is urgently required to replace manual operation. In this paper, an advanced deep-learning method based on machine vision is adopted to detect the surface of corks and classify them automatically. To robustly detect the texture of cork profile, a Multiple Scale Faster R-CNN based on the upper and lower layers (UMS-FRCNN) approach is proposed. The network consists of a region proposal component and the region-of-interest (RoI) detection component. There are two main contributions in our proposed network, which play a significant role in achieving state-of-the-art performance in cork classification. Firstly, the multi-scale information is grouped in two aspects: region suggestion and RoI detection, to detect on small cork texture regions. Secondly, the convolution Region Proposal Network (RPN) is used to generate the candidate regions of the UMS-FRCNN model to improve the classification accuracy. The experimental results show that our method not only can accurately classify the corks but also has better performance than the Faster RCNN. Compared with other methods, our method has the advantages of high precision, less time-consuming, and no dependence labeling. Our work has been successfully utilized in a cork production plant.