Identification Dataset Research Articles

Transmittance Surface Detection and Material Identification Using Multitask ViT-SIFT Fusion

The automatic detection of transmittance surfaces (transparency and translucency) and material class identification (glass, plastic, and so on) from an image is crucial for applications, including domestic service robotics and handling fragile instruments in experimental or industrial settings. However, this is a difficult task due to the lack of expressive textures in such materials. Although previous approaches for detecting transmittance surfaces propose handcrafted features, they disregard global material information and are limited to a single item (usually glass). Material or object recognition algorithms based on convolutional neural networks (CNNs) focus on optimizing performance that requires large labeled datasets. These algorithms are better for detecting opaque objects because large architectures tend to generalize the texture from an image’s background layer. Furthermore, creating large-scale off-the-shelf dataset labeled with transmittance surface is an extremely tedious task. Thus, we propose a simple yet effective and efficient learning model that combines a shallow multitask vision transformer (ViT) with a scale-invariant feature transform (SIFT) and a shallow CNN as the backbone network. SIFT and ViT are used to capture local and global discriminative features, respectively, and their fusion can increase the performance on a small dataset. The backbone network is utilized to introduce the desirable properties of CNN. Using limited training data, this model detects transmittance surfaces and determines their material type. We conducted tests on a new benchmark dataset containing a variety of transmissive materials and three publicly available datasets for robust comparison. Results show that the proposed model achieved a new state-of-the-art accuracy of 81.78% and 77.21% on our new dataset for transmittance surface recognition and material class identification, respectively.

A Novel Method for Few-Shot Specific Emitter Identification in Non-Cooperative Scenarios

Obtaining larger category-label-containing training signal datasets in non-cooperative scenarios is difficult. Moreover, employing smaller labeled signal datasets for specific emitter identification is technically challenging. Therefore, we propose a novel method for few-shot SEI. We first design a bispectral analysis and Radon transformation-based signal preprocessing scheme to obtain feature vectors that effectively characterize the radio frequency fingerprints. The feature vectors are then fed to a network model for feature learning. Moreover, a meta-learning algorithm is applied to the network model to adapt to few-shot feature learning. The conventional meta-learning algorithm is improved to develop a novel algorithm involving latent embedding optimization for meta-learning. The proposed method extracts low-dimensional key features from high-dimensional input data and evaluates the distance and degree of feature dispersion. The resulting information is employed in sample point prediction. The algorithm effectively achieves few-shot SEI and offers stable and efficient recognition after training with a minimum of forty samples. This method identifies emitter individuals under multiple modulation types and exhibits scalability in identifying the emitter numbers. Moreover, it offers adaptability in identifying the emitter individuals under multiple propagation channel types.

AGMG-Net: Leveraging multiscale and fine-grained features for improved cargo recognition.

Security systems place great emphasis on the safety of stored cargo, as any loss or tampering can result in significant economic damage. The cargo identification module within the security system faces the challenge of achieving a 99.99% recognition accuracy. However, current identification methods are limited in accuracy due to the lack of cargo data, insufficient utilization of image features and minimal differences between actual cargo classes. First, we collected and created a cargo identification dataset named "Cargo" using industrial cameras. Subsequently, an Attention-guided Multi-granularity feature fusion model (AGMG-Net) was proposed for cargo identification. This model extracts both coarse-grained and fine-grained features of the cargo using two branch networks and fuses them to fully utilize the information contained in these features. Furthermore, the Attention-guided Multi-stage Attention Accumulation (AMAA) module is introduced for target localization, and the Multi-region Optimal Selection method Based on Confidence (MOSBC) module is used for target cropping. The features from the two branches are fused using a fusion branch in a Concat manner for multi-granularity feature fusion. The experimental results show that the proposed model achieves an average recognition rate of 99.58, 92.73 and 88.57% on the self-built dataset Cargo, and the publicly available datasets Flower and Butterfly20, respectively. This is better than the state-of-the-art model. Therefore, this research method accurately identifies cargo categories and provides valuable assistance to security systems.

Dataset Gambar Wajah untuk Analisis Personal Identification

In today's era, which is supported by technological advances, personal assistance through the face can be carried out by sophisticated machines and robots. One of its applications is personal identification using data mining. But before conducting data training and data classification, it is necessary to carry out the process of data collection and data cleaning or data pre-processing. Currently the face dataset for personal identification at the Caltex Riau Polytechnic in particular is still in the form of raw data, namely in the form of a collection of images that have not been pre-processed. Therefore, this research will perform image preprocessing to clean up the image data that has been collected so that the data can become a cleaner source of information and can be used at a later stage. The data used in this study are image data or photos of Caltex Riau Polytechnic students. At the facial image pre-processing stage using the OpenCV library using the Python programming language. Images collected by 500 students for 5 students. The results of this study are the personal identification dataset of Caltex Riau Polytechnic students consisting of 280 images that have successfully passed the stages of grayscaling, cropping, resizing and Normalization. This dataset is stored in the file data_norm.npz. White box testing is carried out to determine the accuracy of the application of the image pre-processing stage with the test results stating that all functional basis paths applied are in accordance with the cyclometic complexity and its independent path.

Major Migration Channels of Mesoscale Eddies in the Southern Ocean Derived From Satellite Altimetry

AbstractThe propagation pathways of mesoscale eddies have an essential impact on the meridional transport of freshwater, carbon, nutrients, and other tracers. However, existing methods have limitations in exactly expressing their major propagation pathways. To tackle this problem, a direction‐based spatial clustering algorithm for the trajectory is proposed according to the partition‐and‐group framework. The idea is to extract the major eddy channels quantitatively, thus effectively representing the eddy trajectories with similar propagation pathways in a given spatial distance and direction, from the eddy identification and tracking datasets obtained by satellite altimetry spanning nearly 20 years in the Southern Ocean. The major eddy channels of the Southern Ocean are more geographically correlated rather than determined by polarity (anticyclonic eddies (AEs) and cyclonic eddies (CEs)). Importantly, the main axes of zonal propagation for eddy channels are extracted in this study. In the area north of 45°S (30°S–45°S), the zonal propagation (both for AEs and CEs) is intrinsically westward along ∼42°S ± 2° (main axes). While in those south of 45°S (45°S–80°S), it is reversed to eastward, propagating zonally along ∼56°S ± 2° (main axes) and obviously showing a gyre pattern. The bottom topography significantly effects on the propagation pathways of eddy channels in the Southern Ocean. Mainly, their meridional propagation pathways are pronouncedly steered by the shape of oceanic ridges around isobath −3900m (AEs for −3,877 m, CEs for −3,922 m).

A geospatial image based eye movement dataset for cartography and GIS

ABSTRACT Eye movement is a new type of data for cartography and geographic information science (GIS) research. However, previous studies rarely built eye movement datasets with geospatial images. In this paper, we firstly proposed a geospatial image-based eye movement dataset called GeoEye, a publicly shared, widely available eye movement dataset. This dataset consists of 110 college-aged participants who freely viewed 500 images, including thematic maps, remote sensing images, and street view images. In addition, we used the dataset for geospatial image saliency prediction and map user identification. Results demonstrated the scientific benefits and applications of the proposed dataset. GeoEye dataset will not only promote the application of eye-tracking data in cartography and GIS research but also intelligence and customization of geographic information services.

Privacy-preserving integration of multiple institutional data for single-cell type identification with scPrivacy.

The rapid accumulation of large-scale single-cell RNA-seq datasets from multiple institutions presents remarkable opportunities for automatically cell annotations through integrative analyses. However, the privacy issue has existed but being ignored, since we are limited to access and utilize all the reference datasets distributed in different institutions globally due to the prohibited data transmission across institutions by data regulation laws. To this end, we present scPrivacy, which is the first and generalized automatically single-cell type identification prototype to facilitate single cell annotations in a data privacy-preserving collaboration manner. We evaluated scPrivacy on a comprehensive set of publicly available benchmark datasets for single-cell type identification to stimulate the scenario that the reference datasets are rapidly generated and distributed in multiple institutions, while they are prohibited to be integrated directly or exposed to each other due to the data privacy regulations, demonstrating its effectiveness, time efficiency and robustness for privacy-preserving integration of multiple institutional datasets in single cell annotations.

Transformer-based deep learning model and video dataset for unsafe action identification in construction projects

A large proportion of construction accidents are caused by unintentional and unsafe actions and behaviors. It is of significant difficulties and ineffectiveness to monitor unsafe behaviors using conventional manual supervision due to the complex and dynamic working conditions on construction sites. Recently, surveillance videos and computer vision (CV) techniques have been increasingly adopted to automatically identify risky behaviors. However, the challenge remains that spatial and temporal features in video clips cannot be effectively captured and fused by current CV models. To address this challenge, this paper describes a deep learning model named Spatial Temporal Relation Transformer (STR-Transformer), where spatial and temporal features of work behaviors are simultaneously extracted in paralleling video streams and then fused by a specially designed module. To verify the effectiveness of the STR-Transformer, a customized dataset is developed, including seven categories of construction worker behaviors and 1595 video clips. In numerical experiments and case studies, the STR-Transformer achieves an average precision of 88.7%, 4.0% higher than the baseline model. The STR-Transformer enables more accurate and reliable automatic safety surveillance on construction projects, and is expected to reduce accident rates and management costs. Moreover, the performance of STR-Transformer relies on efficient feature integration, which may inspire future studies to identify, extract, and fuse richer features when applying CV-based deep learning models in construction management.

Weed25: A deep learning dataset for weed identification.

Weed suppression is an important factor affecting crop yields. Precise identification of weed species will contribute to automatic weeding by applying proper herbicides, hoeing position determination, and hoeing depth to specific plants as well as reducing crop injury. However, the lack of datasets of weeds in the field has limited the application of deep learning techniques in weed management. In this paper, it presented a dataset of weeds in fields, Weed25, which contained 14,035 images of 25 different weed species. Both monocot and dicot weed image resources were included in this dataset. Meanwhile, weed images at different growth stages were also recorded. Several common deep learning detection models-YOLOv3, YOLOv5, and Faster R-CNN-were applied for weed identification model training using this dataset. The results showed that the average accuracy of detection under the same training parameters were 91.8%, 92.4%, and 92.15% respectively. It presented that Weed25 could be a potential effective training resource for further development of in-field real-time weed identification models. The dataset is available at https://pan.baidu.com/s/1rnUoDm7IxxmX1n1LmtXNXw; the password is rn5h.

The Application of Repeated SMOTE for Multi Class Classification on Imbalanced Data

One of the problems that are often faced by classifier algorithms is related to the problem of imbalanced data. One of the recommended improvement methods at the data level is to balance the number of data in different classes by enlarging the sample to the minority class (oversampling), one of which is called The Synthetic Minority Oversampling Technique (SMOTE). SMOTE is commonly used to balance data consisting of two classes. In this research, SMOTE was used to balance multi-class data. The purpose of this research is to balance multi-class data by applying SMOTE repeatedly. This iterative process needs to be applied if the number of unbalanced data classes is more than two classes, because the one-time SMOTE process is only suitable for binary classification or the number of unbalanced data classes is only one class. To see the performance of iterative SMOTE, the SMOTE datasets were classified using a neural network, k-NN, Nave Bayes, and Random Forest and the performance measures were measured in terms of accuracy, sensitivity, and specificity. The experiment in this research used the Glass Identification dataset which had six classes, and the SMOTE process was repeated five times. The best performance was achieved by the Random Forest classifier method with accuracy = 86.27%, sensitivity = 86.18%, and specificity = 95.82%. The result of experiment present that repeated SMOTE results can increase the performance of classification.

Preparing of ECG Dataset for Biometric ID Identification with Creative Techniques

Preparing of ECG Dataset for Biometric ID Identification with Creative Techniques

Unbiased Identification of Thrombosis Risk Factors in Polycythemia Vera (PV) Using Machine Learning and Rich Data from Automated Extraction of Medical Records Generates Dynamic Models Highly Predictive for Thrombosis in PV

Background: Thrombosis is a leading cause of early morbidity and mortality in patients (pts) with polycythemia vera (PV)1. Thrombosis risk informs treatment decisions in PV but current risk stratification relies upon two non-modifiable factors - Age(</≥ 60) and thrombosis history (hx)2 - thereby offering no potential to monitor changes in risk. We used a rich dataset and machine learning (ML) methods to identify clinically available data useful for dynamic prediction of individualized thrombosis risk in PV. Methods: Under Weill Cornell Medicine (WCM) institutional board approval, we identified in our research repository (RDR) 470 eligible pts with PV diagnosis (Dx) and annotated thrombosis outcomes, as previously described3. We queried electronic medical records using our RDR, Observational Medical Outcomes Partnership, Structured Query Language, and novel workflows implemented in R and Python, to automate extraction of all available clinical data for pts at every clinical visit. Over 2100 clinical parameters were available for ~8400 hematology clinic visits yielding ~1.4 million data elements. These data included all laboratory measures, molecular studies, pathology reports, vitals, medications and clinical hx. Data completeness varied across data types with 283 parameters available in >90% of visits for all pts. In total, 292 parameters with over 50% availability were assessed while others were excluded from analysis. Multivariate imputation by chained equations (MICE) was used to impute missing data. To address potential artifacts from imputation, all analyses were performed on 10 separately imputed datasets and results were aggregated. AutoGluon infrastructure was used to identify the best performing ML algorithms. We then used our dataset to train random forest (RF) ML models that classify, at each clinic visit, the likelihood of thrombosis within a year. Parameters were ranked by Gini feature importance to identify a practical clinical model comprised of a few core parameters (eg 3-7) and interrogated the performance of >2x105 of these models (every combination of top 20 parameters in groups of 3-7) compared to the full model. Model performance was evaluated using the F1 score, which is considered highly predictive when >0.8. Results: Of 470 PV pts, Dx at median age of 54 (range 20-94), 64 (14%) had a thrombosis hx at Dx (8.1% venous, 6.6% arterial). Over a median follow-up of 10 years (yr), 159 thromboses occurred in 115 pts (88 venous, 71 arterial). Cumulative incidence of thrombosis was non-linear, as previously appreciated1. Annual incidence rate (IR) of thrombosis was higher shortly after diagnosis (IR of 4.4% vs 1%, 2yr cutoff), and following a thrombotic event (IR of 9.7% vs 1.8%, 2 yr cutoff). Consistent with this observation, ML models identified time from Dx, and time from prior thrombosis among the top 20 most predictive parameters as assessed by Gini feature importance (Figure 1). Most predictive parameters also included risk factors previously implicated in PV thrombosis (age, blood counts, JAK2 allele burden) as well as known thrombosis risk factors under-recognized in PV (blood type4, BMI, creatinine), and parameters not previously thought to contribute to thrombosis risk (MCV, uric acid, LDH). The full ML model was highly predictive of thrombosis with an F1 of 0.82 (Figure 2). We then identified simplified models (combinations of 4-5 variables) that performed similarly to the full model with an F1 > 0.8. One example combines age, time since Dx, time since prior thrombosis, and BMI. These models are being validated against external datasets and an online risk calculator will be developed. Discussion: We used ML and deep datasets for unbiased identification of the clinical parameters most predictive of near-term thrombosis risk in PV pts. These models address dynamic changes to patient risk due to intrinsic factors (e.g. age, blood type), life events (e.g. time since Dx/thrombosis) and clinical changes (e.g. ANC, BMI) that can be used to tailor risk mitigation strategies. Validation of models with at least core variables of age, time since Dx, time since prior thrombosis, and BMI against external datasets would likely establish a universal model to dynamically predict individualized risk of thrombosis in PV pts. References 1 Hultcrantz. et al. Ann Intern Med. 2018 2 Barbui T, et al.Leukemia 2018 3 Abu-Zeinah G, et al. Leukemia 2021 4 Groot H, et al. Arterioscler Thromb Vasc Biol 2020 Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

A transformer fine-tuning strategy for text dialect identification.

Online medical consultation can significantly improve the efficiency of primary health care. Recently, many online medical question-answer services have been developed that connect the patients with relevant medical consultants based on their questions. Considering the linguistic variety in their question, social background identification of patients can improve the referral system by selecting a medical consultant with a similar social origin for efficient communication. This paper has proposed a novel fine-tuning strategy for the pre-trained transformers to identify the social origin of text authors. When fused with the existing adapter model, the proposed methods achieve an overall accuracy of 53.96% for the Arabic dialect identification task on the Nuanced Arabic Dialect Identification (NADI) dataset. The overall accuracy is 0.54% higher than the previous best for the same dataset, which establishes the utility of custom fine-tuning strategies for pre-trained transformer models.

Player tracking and identification in ice hockey

Tracking and identifying players is a fundamental step in computer vision-based ice hockey analytics. The data generated by tracking is used in many other downstream tasks, such as game event detection and game strategy analysis. Player tracking and identification is a challenging problem since the motion of players in hockey is fast-paced and non-linear when compared to pedestrians. There is also significant camera panning and zooming in hockey broadcast video. Identifying players in ice hockey is challenging since the players of the same team appear almost identical, with the jersey number the only consistent discriminating factor between players. To address this problem, an automated system to track and identify players in broadcast National Hockey League (NHL) videos is introduced. The system is composed of three components (1) player tracking, (2) team identification and (3) player identification. Due to the absence of publicly available datasets, the datasets used to train the three components are annotated manually. Player tracking is performed with the help of a state of the art tracking algorithm obtaining a Multi-Object Tracking Accuracy (MOTA) score of 94.5%. For team identification, away-team jerseys are grouped into a single class and home-team jerseys are grouped in classes according to their jersey color. A convolutional neural network is then trained on the team identification dataset. The team identification network obtains an accuracy of 97% on the test set. A novel player identification model is introduced that utilizes a temporal one-dimensional convolutional network to identify players from player bounding box sequences. The player identification model further takes advantage of the available NHL game roster data to obtain a player identification accuracy of 83%.

A deep learning loss based on additive cosine margin: Application to fashion style and face recognition

Recently, loss functions based on angular spans improved the performance of deep visual recognition. These losses converted Euclidean cross entropy to angular cross entropy loss. Fashion style recognition deals with the problem of assigning a person’s outfit to a fashion style category. Due to the high similarity between different clothing items and the use of softmax-based loss functions, many of the current methods that address this problem show relatively poor performance and cannot guarantee sufficient inter-class margins in the fashion domain. In this work, we propose an end-to-end method for deep visual recognition by combining a standard CNN architecture with a novel loss function, which we call Additive Cosine Margin Loss (ACML). The proposed function not only projects feature vectors of different classes into different regions of the embedding, but also enforces compactness of the projections within each class. Our experiments were conducted on two public and well-known fashion style recognition datasets FashionStyle14 and HipsterWars, and on the face verification and identification datasets LFW, YTF, and MegaFace. These experiments demonstrate the superiority of the proposed loss function over: (1) existing angular margin-based loss functions (2) state-of-the-art methods for clothing style recognition as well as face analysis tasks.

Simplifying VGG-16 for Plant Species Identification

Plant species identification represents an extraordinary challenge for machine learning due to visual interspecies similarities and large intraspecies variations. Furthermore, research literature reports that plant species identification usually lacks sufficiently large datasets for training classification models. In this paper, we address this problem with a model that simplifies the VGG-16 architecture, the N-VGG model. The idea behind N-VGG is to reduce experimentally observed overfitting on VGG-16 by using as few trainable parameters as possible. To do this, we substitute the flattening layer on the VGG architecture with a global average pooling layer. This reduces the size of the feature vector. In addition, we eliminate one of the two fully-connected layers and use a new hyper-parameter, N, to indicate the number of nodes on the remaining layer. To show the robustness of the N-VGG model, we conducted extensive experimentation. We trained N-VGG on five datasets for plant species identification. Four of these datasets are publicly available and have been widely used as benchmarks for plant identification models. For all datasets, we compare the accuracy of N-VGG to that of the VGG-16, Inception-v4, and EfficienNet-B3 models. The experimental results show that the N-VGG model achieved the best classification performance for all but one datasets, whereas all the models showed a remarkable performance for the remaining dataset. This evidence supports our initial idea that, for plant species classification, some accuracy might be lost due to overfitting and that having fewer trainable parameters helps in producing a more robust model.

User recognition in participatory sensing systems using deep learning based on spectro-temporal representation of accelerometer signals

Identification of the actual sensor data provider is important in smartphone-based participatory sensing systems. More specifically, the participatory sensing systems must be able to identify the actual users from smartphone-embedded sensors’ data in a fast, silent, and unobtrusive manner, ensuring that appropriate privacy protection and authentication are provided automatically. In this context, this paper investigates smartphone-equipped accelerometer signal data to determine a person’s distinctive gait, allowing different providers to be recognized. The problem of user identification is formulated as an image classification task by changing the accelerometer signal into a 2D spectrotemporal image representation by calculating Short-Term Fourier Transform (STFT) of the signal. A deep learning based custom CNN model is proposed and evaluated on real-world benchmark dataset (containing 60 million accelerometer data samples of 387 users) for user identification. The efficacy of the proposed deep learning based approach is verified by comparing its performance with several baseline methods including Multi-layer Perceptron, K-Nearest Neighbours, Logistic Regression, Random Forest, and Decision Tree. The results of the experiments confirm that the proposed technique yields better performance (accuracy = 0.98, precision = 0.96, recall = 0.98, F1-score = 0.96, AUC = .9995) with respect to the baselines. Furthermore, the proposed model outperforms other CNN architectures (e.g., Lenet-5, AlexNet, and ZFNet) across all performance metrics.

The Good Sleeper Scale-15 items: a questionnaire for the standardised assessment of good sleepers.

Research with 'good sleepers' is ubiquitous, yet there are no standardised criteria to identify a 'good sleeper'. The present study aimed to create and validate a questionnaire for identifying good sleepers for use in research studies known as the Good Sleeper Scale-15 items (GSS-15). Data were derived from a population-based survey of Australian adults (n=2,044). A total of 23 items were chosen for possible inclusion. An exploratory factor analysis (EFA) was conducted on ~10% of the survey dataset (n =191) for factor identification and item reduction. A confirmatory factor analysis (CFA) was conducted on the remaining data (n=1,853) to test model fit. Receiver operating characteristic curves and correlations were conducted to derive cut-off scores and test associations with sleep, daytime functioning, health, and quality-of-life. The EFA identified six factors: 'Sleep Difficulties', 'Timing', 'Duration', 'Regularity', 'Adequacy', and 'Perceived Sleep Problem'. The CFA showed that model fit was high and comparable to other sleep instruments, χ2 (63)=378.22, p < 0.001, root mean square error of approximation=0.05, with acceptable internal consistency (α=0.76). Strong correlations were consistently found between GSS-15 global scores and outcomes, including 'a good night's sleep' (r =0.7), 'feeling un-refreshed' (r = -0.59), and 'experienced sleepiness' (r = -0.51), p < 0.001. Cut-off scores were derived to categorise individuals likely to be a good sleeper (GSS-15 score ≥40) and those very likely to be a good sleeper (GSS-15 score ≥45). The GSS-15 is a freely available, robust questionnaire that will assist in identifying good sleepers for the purpose of sleep research. Future work will test relationships with other sleep measures in community and clinical samples.

Classification Method of Significant Rice Pests Based on Deep Learning

Rice pests are one of the main factors affecting rice yield. The accurate identification of pests facilitates timely preventive measures to avoid economic losses. Some existing open source datasets related to rice pest identification mostly include only a small number of samples, or suffer from inter-class and intra-class variance and data imbalance challenges, which limit the application of deep learning techniques in the field of rice pest identification. In this paper, based on the IP102 dataset, we first reorganized a large-scale dataset for rice pest identification by Web crawler technique and manual screening. This dataset was given the name IP_RicePests. Specifically, the dataset includes 8248 images belonging to 14 categories. The IP_RicePests dataset was then expanded to include 14,000 images via ARGAN data augmentation technique to address the difficulties in obtaining large samples of rice pests. Finally, the parameters trained on the public image ImageNet dataset using VGGNet, ResNet and MobileNet networks were used as the initial values of the target data training network to achieve image classification in the field of rice pests. The experimental results show that all three classification networks combined with transfer learning have good recognition accuracy, among which the highest classification accuracy can be obtained on the IP_RicePests dataset via fine-tuning the parameters of the VGG16 network. In addition, following ARGAN data augmentation the dataset demonstrates high accuracy improvements in all three models, and fine-tuning the VGG16 network parameters obtains the highest accuracy in the augmented IP_RicePests dataset. It is demonstrated that CNN combined with transfer learning can employ the ARGAN data augmentation technique to overcome difficulties in obtaining large sample sizes and improve the efficiency of rice pest identification. This study provides foundational data and technical support for rice pest identification.

Image Information Contribution Evaluation for Plant Diseases Classification via Inter-Class Similarity

Combineingplant diseases identification and deep learning algorithm can achieve cost-effective prevention effect, and has been widely used. However, the current field of intelligent plant diseases identification still faces the problems of insufficient data and inaccurate classification. Aiming to resolve these problems, the present research proposes an image information contribution evaluation method based on the analysis of inter-class similarity. Combining this method with the active learning image selection strategy can provide guidance for the collection and annotation of intelligent identification datasets of plant diseases, so as to improve the recognition effect and reduce the cost. The method proposed includes two modules: the inter-classes similarity evaluation module and the image information contribution evaluation module. The images located on the decision boundary between high similarity classes will be analysis as high information contribution images, they will provide more information for plant diseases classification. In order to verify the effectiveness of this method, experiments were carried on the fine-grained classification dataset of tomato diseases. Experimental results confirm the superiority of this method compared with others. This research is in the field of plant disease classification. For the detection and segmentation, further research is advisable.