Leaf Segmentation and Classification with a Complicated Background Using Deep Learning

The segmentation and classification of leaves in plant images are a great challenge, especially when several leaves are overlapping in images with a complicated background. In this paper, the segmentation and classification of leaf images with a complicated background using deep learning are studied. First, more than 2500 leaf images with a complicated background are collected and artificially labeled with target pixels and background pixels. Two-thousand of them are fed into a Mask Region-based Convolutional Neural Network (Mask R-CNN) to train a model for leaf segmentation. Then, a training set that contains more than 1500 training images of 15 species is fed into a very deep convolutional network with 16 layers (VGG16) to train a model for leaf classification. The best hyperparameters for these methods are found by comparing a variety of parameter combinations. The results show that the average Misclassification Error (ME) of 80 test images using Mask R-CNN is 1.15%. The average accuracy value for the leaf classification of 150 test images using VGG16 is up to 91.5%. This indicates that these methods can be used to segment and classify the leaf image with a complicated background effectively. It could provide a reference for the phenotype analysis and automatic classification of plants.

Download Full-text

Zebrafish Larvae Phenotype Classification from Bright-field Microscopic Images Using a Two-Tier Deep-Learning Pipeline

Applied Sciences ◽

10.3390/app10041247 ◽

2020 ◽

Vol 10 (4) ◽

pp. 1247 ◽

Cited By ~ 3

Author(s):

Shang Shang ◽

Sijie Lin ◽

Fengyu Cong

Keyword(s):

Deep Learning ◽

Environmental Influence ◽

Training Data ◽

Convolutional Network ◽

Training Images ◽

Zebrafish Larvae ◽

Machine Learning Methods ◽

Phenotype Classification ◽

Convolution Kernels

Classification of different zebrafish larvae phenotypes is useful for studying the environmental influence on embryo development. However, the scarcity of well-annotated training images and fuzzy inter-phenotype differences hamper the application of machine-learning methods in phenotype classification. This study develops a deep-learning approach to address these challenging problems. A convolutional network model with compressed separable convolution kernels is adopted to address the overfitting issue caused by insufficient training data. A two-tier classification pipeline is designed to improve the classification accuracy based on fuzzy phenotype features. Our method achieved an averaged accuracy of 91% for all the phenotypes and maximum accuracy of 100% for some phenotypes (e.g., dead and chorion). We also compared our method with the state-of-the-art methods based on the same dataset. Our method obtained dramatic accuracy improvement up to 22% against the existing method. This study offers an effective deep-learning solution for classifying difficult zebrafish larvae phenotypes based on very limited training data.

Download Full-text

Qualitative classification of waste textiles based on near infrared spectroscopy and the convolutional network

Textile Research Journal ◽

10.1177/0040517519886032 ◽

2019 ◽

Vol 90 (9-10) ◽

pp. 1057-1066 ◽

Cited By ~ 1

Author(s):

Zhengdong Liu ◽

Wenxia Li ◽

Zihan Wei

Keyword(s):

Deep Learning ◽

Infrared Spectroscopy ◽

Near Infrared Spectroscopy ◽

Near Infrared ◽

Convolutional Network ◽

The Sustainable Development ◽

Softmax Classifier ◽

Qualitative Classification ◽

Reliable Classification

The recycling of waste textiles has become a growth point for the sustainable development of the textile and clothing industry. In addition, sorting is a key link in the follow-up recycling process. Since different fabrics are required to be processed by different technologies, manual sorting not only takes time and effort but also cannot achieve accurate and reliable classification. Based on the analysis of near infrared spectroscopy, the theory and methods of deep learning are used for the qualitative classification of waste textiles in order to complete the automatic fabric composition recognition in the sorting process. Firstly, a standard sample set is established by waveform clipping and normalization, and a Textile Recycling Net deep web suitable for near infrared spectroscopy is established. Then, a pixilated layer is used to facilitate the deep learning of features, and the multidimensional features of the spectrum are extracted by using the multi-layer convolutional and pooling layers. Finally, the softmax classifier is adopted to complete the qualitative classification. Experimental results show that the convolutional network classification method using normalized and pixelated near infrared spectroscopy can realize the automatic classification of several common textiles, such as cotton and polyester, and effectively improve the detection level and speed of fabric components.

Download Full-text

Modern Approaches to Detect and Classify Comment Toxicity Using Neural Networks

Modeling and Analysis of Information Systems ◽

10.18255/1818-1015-2020-1-48-61 ◽

2020 ◽

Vol 27 (1) ◽

pp. 48-61 ◽

Cited By ~ 1

Author(s):

Sergey V. Morzhov

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Language Processing ◽

Rapid Evolution ◽

Learning Technologies ◽

Training Set ◽

Online Platforms ◽

Processing Algorithms ◽

New Algorithms

The growth of popularity of online platforms which allow users to communicate with each other, share opinions about various events, and leave comments boosted the development of natural language processing algorithms. Tens of millions of messages per day are published by users of a particular social network need to be analyzed in real time for moderation in order to prevent the spread of various illegal or offensive information, threats and other types of toxic comments. Of course, such a large amount of information can be processed quite quickly only automatically. that is why there is a need to and a way to teach computers to “understand” a text written by humans. It is a non-trivial task even if the word “understand” here means only “to classify”. the rapid evolution of machine learning technologies has led to ubiquitous implementation of new algorithms. A lot of tasks, which for many years were considered almost impossible to solve, are now quite successfully solved using deep learning technologies. this article considers algorithms built using deep learning technologies and neural networks which can successfully solve the problem of detection and classification of toxic comments. In addition, the article presents the results of the developed algorithms, as well as the results of the ensemble of all considered algorithms on a large training set collected and tagged by Google and Jigsaw.

Download Full-text

Classification of the Trap-Neuter-Return Surgery Images of Stray Animals Using Yolo-Based Deep Learning Integrated with a Majority Voting System

Applied Sciences ◽

10.3390/app11188578 ◽

2021 ◽

Vol 11 (18) ◽

pp. 8578

Author(s):

Yi-Cheng Huang ◽

Ting-Hsueh Chuang ◽

Yeong-Lin Lai

Keyword(s):

Deep Learning ◽

Majority Voting ◽

Effective Solution ◽

Stray Cats ◽

Average Accuracy ◽

Voting System ◽

The Cost ◽

Animal Care ◽

Manual Review

Trap-neuter-return (TNR) has become an effective solution to reduce the prevalence of stray animals. Due to the non-culling policy for stray cats and dogs since 2017, there is a great demand for the sterilization of cats and dogs in Taiwan. In 2020, Heart of Taiwan Animal Care (HOTAC) had more than 32,000 cases of neutered cats and dogs. HOTAC needs to take pictures to record the ears and excised organs of each neutered cat or dog from different veterinary hospitals. The correctness of the archived medical photos and the different shooting and imaging angles from different veterinary hospitals must be carefully reviewed by human professionals. To reduce the cost of manual review, Yolo’s ensemble learning based on deep learning and a majority voting system can effectively identify TNR surgical images, save 80% of the labor force, and its average accuracy (mAP) exceeds 90%. The best feature extraction based on the Yolo model is Yolov4, whose mAP reaches 91.99%, and the result is integrated into the voting classification. Experimental results show that compared with the previous manual work, it can decrease the workload by more than 80%.

Download Full-text

Classification of Stages Diabetic Retinopathy Using MobileNetV2 Model

10.29007/h46n ◽

2022 ◽

Author(s):

Hoang Nhut Huynh ◽

Minh Thanh Do ◽

Gia Thinh Huynh ◽

Anh Tu Tran ◽

Trung Nghia Tran

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

High Performance ◽

Vision Loss ◽

Training Model ◽

Average Accuracy ◽

Common Diagnosis ◽

Learning Technique ◽

The Common

Diabetic retinopathy (DR) is a complication of diabetes mellitus that causes retinal damage that can lead to vision loss if not detected and treated promptly. The common diagnosis stages of the disease take time, effort, and cost and can be misdiagnosed. In the recent period with the explosion of artificial intelligence, deep learning has become the most popular tool with high performance in many fields, especially in the analysis and classification of medical images. The Convolutional Neural Network (CNN) is more widely used as a deep learning method in medical imaging analysis with highly effective. In this paper, the five-stage image of modern DR (healthy, mild, moderate, severe, and proliferative) can be detected and classified using the deep learning technique. After cross-validation training and testing on the corresponding 5,590-image dataset, a pre-MobileNetV2 training model is proposed in classifying stages of diabetic retinopathy. The average accuracy of the model achieved was 93.89% with the precision of 94.00%, recall 92.00% and f1-score 90.00%. The corresponding thermal image is also given to help experts for evaluating the influence of the retina in each different stage.

Download Full-text

Classification of EgyptSat-1 Images Using Deep Learning Methods

International Journal of Sensors Wireless Communications and Control ◽

10.2174/2210327909666190207153858 ◽

2020 ◽

Vol 10 (1) ◽

pp. 37-46 ◽

Cited By ~ 3

Author(s):

Hatem Keshk ◽

Xu-Cheng Yin

Keyword(s):

Neural Network ◽

Remote Sensing ◽

Deep Learning ◽

Convolutional Neural Network ◽

High Performance ◽

Satellite Images ◽

Satellite Image ◽

Aerial Photographs ◽

Average Accuracy

Background: Deep Learning (DL) neural network methods have become a hotspot subject of research in the remote sensing field. Classification of aerial satellite images depends on spectral content, which is a challenging topic in remote sensing. Objective: With the aim to accomplish a high performance and accuracy of Egyptsat-1 satellite image classification, the use of the Convolutional Neural Network (CNN) is raised in this paper because CNN is considered a leading deep learning method. CNN is developed to classify aerial photographs into land cover classes such as urban, vegetation, desert, water bodies, soil, roads, etc. In our work, a comparison between MAXIMUM Likelihood (ML) which represents the traditional supervised classification methods and CNN method is conducted. Conclusion: This research finds that CNN outperforms ML by 9%. The convolutional neural network has better classification result, which reached 92.25% as its average accuracy. Also, the experiments showed that the convolutional neural network is the most satisfactory and effective classification method applied to classify Egyptsat-1 satellite images.

Download Full-text

Traditional Bangladeshi Sports Video Classification Using Deep Learning Method

Applied Sciences ◽

10.3390/app11052149 ◽

2021 ◽

Vol 11 (5) ◽

pp. 2149

Author(s):

Moumita Sen Sarma ◽

Kaushik Deb ◽

Pranab Kumar Dhar ◽

Takeshi Koshiba

Keyword(s):

Deep Learning ◽

Short Term Memory ◽

Easy Access ◽

Learning Method ◽

Cultural Significance ◽

Sports Video ◽

Temporal Features ◽

Average Accuracy ◽

Proposed Model

Sports activities play a crucial role in preserving our health and mind. Due to the rapid growth of sports video repositories, automatized classification has become essential for easy access and retrieval, content-based recommendations, contextual advertising, etc. Traditional Bangladeshi sport is a genre of sports that bears the cultural significance of Bangladesh. Classification of this genre can act as a catalyst in reviving their lost dignity. In this paper, the Deep Learning method is utilized to classify traditional Bangladeshi sports videos by extracting both the spatial and temporal features from the videos. In this regard, a new Traditional Bangladeshi Sports Video (TBSV) dataset is constructed containing five classes: Boli Khela, Kabaddi, Lathi Khela, Kho Kho, and Nouka Baich. A key contribution of this paper is to develop a scratch model by incorporating the two most prominent deep learning algorithms: convolutional neural network (CNN) and long short term memory (LSTM). Moreover, the transfer learning approach with the fine-tuned VGG19 and LSTM is used for TBSV classification. Furthermore, the proposed model is assessed over four challenging datasets: KTH, UCF-11, UCF-101, and UCF Sports. This model outperforms some recent works on these datasets while showing 99% average accuracy on the TBSV dataset.

Download Full-text

Alternately Updated Spectral–Spatial Convolution Network for the Classification of Hyperspectral Images

Remote Sensing ◽

10.3390/rs11151794 ◽

2019 ◽

Vol 11 (15) ◽

pp. 1794 ◽

Cited By ~ 3

Author(s):

Wenju Wang ◽

Shuguang Dou ◽

Sen Wang

Keyword(s):

Deep Learning ◽

Forward Direction ◽

Hyperspectral Images ◽

Data Sets ◽

Convolutional Network ◽

Spatial Features ◽

Training Samples ◽

Feedback Structure ◽

Kennedy Space

The connection structure in the convolutional layers of most deep learning-based algorithms used for the classification of hyperspectral images (HSIs) has typically been in the forward direction. In this study, an end-to-end alternately updated spectral–spatial convolutional network (AUSSC) with a recurrent feedback structure is used to learn refined spectral and spatial features for HSI classification. The proposed AUSSC includes alternating updated blocks in which each layer serves as both an input and an output for the other layers. The AUSSC can refine spectral and spatial features many times under fixed parameters. A center loss function is introduced as an auxiliary objective function to improve the discrimination of features acquired by the model. Additionally, the AUSSC utilizes smaller convolutional kernels than other convolutional neural network (CNN)-based methods to reduce the number of parameters and alleviate overfitting. The proposed method was implemented on four HSI data sets, as follows: Indian Pines, Kennedy Space Center, Salinas Scene, and Houston. Experimental results demonstrated that the proposed AUSSC outperformed the HSI classification accuracy obtained by state-of-the-art deep learning-based methods with a small number of training samples.

Download Full-text

Tens of images can suffice to train neural networks for malignant leukocyte detection

Scientific Reports ◽

10.1038/s41598-021-86995-5 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Jens P. E. Schouten ◽

Christian Matek ◽

Luuk F. P. Jacobs ◽

Michèle C. Buck ◽

Dragan Bošnački ◽

...

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Single Cell ◽

Expert Knowledge ◽

Binary Classification ◽

Lymphoblastic Leukemia ◽

Training Set ◽

Early Detection Of Disease ◽

Set Size ◽

Training Images

AbstractConvolutional neural networks (CNNs) excel as powerful tools for biomedical image classification. It is commonly assumed that training CNNs requires large amounts of annotated data. This is a bottleneck in many medical applications where annotation relies on expert knowledge. Here, we analyze the binary classification performance of a CNN on two independent cytomorphology datasets as a function of training set size. Specifically, we train a sequential model to discriminate non-malignant leukocytes from blast cells, whose appearance in the peripheral blood is a hallmark of leukemia. We systematically vary training set size, finding that tens of training images suffice for a binary classification with an ROC-AUC over 90%. Saliency maps and layer-wise relevance propagation visualizations suggest that the network learns to increasingly focus on nuclear structures of leukocytes as the number of training images is increased. A low dimensional tSNE representation reveals that while the two classes are separated already for a few training images, the distinction between the classes becomes clearer when more training images are used. To evaluate the performance in a multi-class problem, we annotated single-cell images from a acute lymphoblastic leukemia dataset into six different hematopoietic classes. Multi-class prediction suggests that also here few single-cell images suffice if differences between morphological classes are large enough. The incorporation of deep learning algorithms into clinical practice has the potential to reduce variability and cost, democratize usage of expertise, and allow for early detection of disease onset and relapse. Our approach evaluates the performance of a deep learning based cytology classifier with respect to size and complexity of the training data and the classification task.

Download Full-text

Acral melanoma detection using dermoscopic images and convolutional neural networks

Visual Computing for Industry Biomedicine and Art ◽

10.1186/s42492-021-00091-z ◽

2021 ◽

Vol 4 (1) ◽

Author(s):

Qaiser Abbas ◽

Farheen Ramzan ◽

Muhammad Usman Ghani

Keyword(s):

Deep Learning ◽

Skin Cancer ◽

Transfer Learning ◽

Clinical Decision Making ◽

Data Augmentation ◽

Automated System ◽

Convolutional Network ◽

Acral Melanoma ◽

Melanoma Subtypes

AbstractAcral melanoma (AM) is a rare and lethal type of skin cancer. It can be diagnosed by expert dermatologists, using dermoscopic imaging. It is challenging for dermatologists to diagnose melanoma because of the very minor differences between melanoma and non-melanoma cancers. Most of the research on skin cancer diagnosis is related to the binary classification of lesions into melanoma and non-melanoma. However, to date, limited research has been conducted on the classification of melanoma subtypes. The current study investigated the effectiveness of dermoscopy and deep learning in classifying melanoma subtypes, such as, AM. In this study, we present a novel deep learning model, developed to classify skin cancer. We utilized a dermoscopic image dataset from the Yonsei University Health System South Korea for the classification of skin lesions. Various image processing and data augmentation techniques have been applied to develop a robust automated system for AM detection. Our custom-built model is a seven-layered deep convolutional network that was trained from scratch. Additionally, transfer learning was utilized to compare the performance of our model, where AlexNet and ResNet-18 were modified, fine-tuned, and trained on the same dataset. We achieved improved results from our proposed model with an accuracy of more than 90 % for AM and benign nevus, respectively. Additionally, using the transfer learning approach, we achieved an average accuracy of nearly 97 %, which is comparable to that of state-of-the-art methods. From our analysis and results, we found that our model performed well and was able to effectively classify skin cancer. Our results show that the proposed system can be used by dermatologists in the clinical decision-making process for the early diagnosis of AM.

Download Full-text