Deep Learning Based Hand Detection in Cluttered Environment Using Skin Segmentation

Roy, Kankana; Mohanty, Amiya Ranjan; Sahay, Rajiv R.

doi:10.1109/iccvw.2017.81

Cited by 53 publications

(46 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have also used the same dataset for training in all our experiments to provide an unbiased comparison. To assess the detector performance and demonstrate the robustness and the generalization power of the hand detector, we evaluated the performance of the trained models on four different test datasets, namely, the Oxford [24], 5-signers [45], EgoHands [44], and Indian classical dance (ICD) datasets ([5]). Table 2 summarizes the characteristics of the datasets used for the training and testing of the hand detector.Hand gesture recognition datasets: To train and evaluate the gesture recognition performance of our proposed architecture, two hand gesture datasets were chosen because they both have publicly available data with challenging data conditions, i.e., they contain a large amount of data with a different number of classes, as detailed below.…”

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Experiments and Discussionmentioning

confidence: 99%

“…Combining skin detection with deep learning object detectors has also been proposed. Roy et al [5] proposed two architectures (patch-based CNN and regression-based CNN for skin segmentation). Their main purpose was to reduce the occurrence of false positives resulting from the estimated bounding boxes of recent object detectors [14,15].…”

Section: Related Workmentioning

confidence: 99%

“…Consequently, these methods have been developed and used for hand detection. Roy et al [5] proposed a two-stage hand detector based on the region-CNN (R-CNN) and Faster R-CNN [14,15] frameworks. Initially, they used an object detection algorithm to generate hand regions and then a CNN-based skin segmentation was used to reduce occurrences of false positives during hand detection.…”

Section: Related Workmentioning

confidence: 99%

“…This task presents a challenging problem that has not yet been solved in computer vision and machine learning. Unlike many previous studies [1,2,3,4,5,6,7], which separately have tried to address hand detection or hand gesture recognition, our approach attempts to jointly solve the problem of hand localization and gesture recognition. This task, however, is very challenging, due to the significant variations of hand images in realistic scenarios.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Deep Learning-Based End-to-End Composite System for Hand Detection and Gesture Recognition

Mohammed

Islam

2019

Sensors

View full text Add to dashboard Cite

Recent research on hand detection and gesture recognition has attracted increasing interest due to its broad range of potential applications, such as human-computer interaction, sign language recognition, hand action analysis, driver hand behavior monitoring, and virtual reality. In recent years, several approaches have been proposed with the aim of developing a robust algorithm which functions in complex and cluttered environments. Although several researchers have addressed this challenging problem, a robust system is still elusive. Therefore, we propose a deep learning-based architecture to jointly detect and classify hand gestures. In the proposed architecture, the whole image is passed through a one-stage dense object detector to extract hand regions, which, in turn, pass through a lightweight convolutional neural network (CNN) for hand gesture recognition. To evaluate our approach, we conducted extensive experiments on four publicly available datasets for hand detection, including the Oxford, 5-signers, EgoHands, and Indian classical dance (ICD) datasets, along with two hand gesture datasets with different gesture vocabularies for hand gesture recognition, namely, the LaRED and TinyHands datasets. Here, experimental results demonstrate that the proposed architecture is efficient and robust. In addition, it outperforms other approaches in both the hand detection and gesture classification tasks.

show abstract

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Deep Learning-Based End-to-End Composite System for Hand Detection and Gesture Recognition

Mohammed

Islam

2019

Sensors

View full text Add to dashboard Cite

show abstract

Diagnosis of Alternaria disease and leafminer pest on tomato leaves using image processing techniques

2022

View full text Add to dashboard Cite

BACKGROUND: Diseases such as Alternaria and pests such as leafminer threaten tomato as one of the most widely used agricultural products. These pests and diseases first damage the leaves of tomatoes, then the flowers, and finally the fruit. Therefore, the damage to the tomato tree must be controlled in its early stages. It is difficult for farmers to distinguish Alternaria disease from leafminer pest at the early and middle stages of their outbreak on tomato leaves. In the present study, 272 tomato leaf images were prepared from the farm of the Vali-e-Asr University of Rafsanjan, including 100 healthy leaves and 172 infected leaves with both Alternaria and leafminer at the initial stages. The image processing technique, texture, neural networks and adaptive network-based fuzzy inference system (ANFIS) classifiers were used to diagnose Alternaria disease and leafminer pest on this dataset. RESULTS:The results showed that the ANFIS classifier achieved an accuracy of 84.71% when performing an equal error rate, 87.78% in the area under the curve, and 85.23% in 3.26 s on the central processing unit for the segmentation of Alternaria disease and leafminer pest in RGB color space. Also, the accuracy of 90% and 98% were obtained for segmentation and classification on the PlantVillage dataset in YCBCR color space.CONCLUSION: The present study suggests a high classification accuracy for an intelligent selection of pixel values to train the ANFIS classifier. This classifier has high accuracy and speed, low sensitivity to the light intensity of images, and practical application in diagnosing various diseases and pests on numerous datasets.

show abstract

Fast Detection and Rule Based Classification of Bharatanatyam hasta mudra

Paul

Das

2023

Communications in Computer and Information Science

View full text Add to dashboard Cite

Deep Learning Based Hand Detection in Cluttered Environment Using Skin Segmentation

Cited by 53 publications

References 35 publications

A Deep Learning-Based End-to-End Composite System for Hand Detection and Gesture Recognition

A Deep Learning-Based End-to-End Composite System for Hand Detection and Gesture Recognition

Diagnosis of Alternaria disease and leafminer pest on tomato leaves using image processing techniques

Fast Detection and Rule Based Classification of Bharatanatyam hasta mudra

Contact Info

Product

Resources

About