Papers by Mahasweta Joshi
Face recognition has been important research in the field of computer vision technology for a ver... more Face recognition has been important research in the field of computer vision technology for a very long time. It is an efficient and one of the most preferred biometric technique, The proposed algorithms in this field are endless, and the accuracy that can be achieved higher than ever due to vast advancement in deep learning technology, video-based facial recognition it plays a major role in real-life applications but there are some challenges also, Deep learning techniques archives state-of-the-art results in the field of computer vision, At present deep learning techniques are one of the dominant technique compare to other technique, This paper presents a comprehensive survey on the different issues related to Video-based face recognition systems also to solving these issues are analyzed by presenting existing techniques that have been proposed in the literature and various Real-world applications of video-based face recognition.
Turkish Online Journal of Qualitative Inquiry, Aug 4, 2021
Detecting and segmenting Salient and co-salient object in an image and video has attracted a lot ... more Detecting and segmenting Salient and co-salient object in an image and video has attracted a lot of interest in computer vision. Now a day's many applications have emerged and models have been proposed. We aim to supply a comprehensive analysis of recent add salient object detection and provides this field, completely different areas like generic scene segmentation, object proposal generation, strikingness for fixation prediction and co-saliency detection via looking deep and wide. Covering (20) publications, our survey (1) root, key concepts and task (2) core techniques and modeling trends, and (3) datasets that are used in the salient object detection and segmentation. We also suggest future research work. Index Terms Salient object detection, bottom-up saliency, explicit saliency, visual attention, a region of interest.
A face Recognition method is implemented to find a particular person. Here we are detected specif... more A face Recognition method is implemented to find a particular person. Here we are detected specific player from whole cricket inning video and summarize it, there are many techniques are available for face detection, in this paper we are discussed some of them in brief. Index Terms Object Detect Using SURF, Face Recognition Using PCA, LDA.
In this paper we study and represent vari ous techniques of image super resolution (SR), it i s a... more In this paper we study and represent vari ous techniques of image super resolution (SR), it i s also called as super resolution reconstruction. Super re solution reconstruction produces high resolution im age from sequence of low resolution images. The main aim of super resolution is to improve visual quality of available low resolution image. Also existing Low Resolution (LR) imaging can be utilized with help of super resolution reconstruction. This paper focuses on re construction based and learning based super resolution methods. We have represented various existing super resolution techniques, advantages and disadvantages of those techniques, recent work and recent methods of super resolution reconstruction method. Finally we have presented challenge issues and future research directions for super resolution.
Journal of emerging technologies and innovative research, 2018
From the ancient time, women have made marvelous contributions to the development of their respec... more From the ancient time, women have made marvelous contributions to the development of their respective societies. These are vary according to each successive stage of human development. They contribute extremely well in different stages of life ranging from child birth, moral upbringing, superior health and hygiene, nutritional values, advice to men as mothers, wives, sisters and kinswomen, religious, socio-economic and political developments. That’s why, women are described as the heart-beat of the society, because almost nothing can be achieved without their impact either directly or indirectly in the concerns related to men. This paper contains importance of education, barriers for getting it, significance of female education, social importance and hurdles of women’s after being educated.
Image Segmentation is an important image processing technique which is used to analyse colou r, t... more Image Segmentation is an important image processing technique which is used to analyse colou r, texture etc. Image Segmentation is used to separate an image into several “meaningful” parts. Normaliz ed cut (Ncut) is based on graph cut technique to solve the image Segmentation problems. Rather than just focusing on local features and their consistencies, Ncut consider the global impression of an image. W e have applied Ncut algorithm, on many images and successf ully segmented the images into meaningful parts. Key Terms: Normalized cut (Ncut); Active contour model (snake); mean shift image segmentation
Sign Language is most accepted and meaningful way of communication for deaf and dumb people of th... more Sign Language is most accepted and meaningful way of communication for deaf and dumb people of the society. Sign language uses gestures, head/body movements and facial expressions for communication. It is a powerful means of Non-Verbal communication among humans. Every country has its own developed Sign Language. The language which is used in India is called as “Indian Sign Language (ISL)”. Only little research work has been carried out in this area as ISL has got standardized recently. Currently many researchers have mainly focused on gesture recognition that has been recorded under static hand gesture. Only few works have been reported for recognition of dynamic hand gesture. Many methods have been developed to recognize alphabets and numerals of ISL. The major steps involved in designing such a system are: gesture acquisition, tracking and segmentation, feature extraction and gesture recognition. This paper presents a survey on various hand gesture recognition approaches for ISL.
Abnormal crowd behavior detection is an important part in video surveillance system because of mo... more Abnormal crowd behavior detection is an important part in video surveillance system because of more concern with people safety in public places like sports stadium, shopping mall, train subway etc. Crowd congestion is not fixed in public places so due to heavy density and occlusion abnormal crowd behavior detection become challenging task. In crowd segmentation and tracking of every person is not possible and even it’s become difficult task so we require method which does not require tracking of every person individually. This paper presents several techniques for abnormal behavior Detection such as crowd running suddenly, congestion detection, fighting etc. Keyword: Abnormal Crowd Behavior, Video Surveillance, Optical Flow, Spatio Temporal Feature, crowd motion.
Nowadays visual information is gaining much more importance than textual information as pictures ... more Nowadays visual information is gaining much more importance than textual information as pictures convey more than words. As image has richer content than text, Video database is increasing exponentially over the Internet. Content-based Video retrieval (CBVR) from large video storage has become a challenging and important issue for researchers over worldwide. There is lack of tools so it is difficult to distinguish between low and high level feature. There has been lot of CBVR systems developed but still there is a lack of efficient retrieval system, as many retrieval techniques are not suitable for extensive search and so it deteriorates the overall performance of system. In this paper we have given an outline of the sequential steps of existing CBVR systems.
2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS), 2017
Sign language is an accepted language for communication between deaf and dumb community people. I... more Sign language is an accepted language for communication between deaf and dumb community people. It is the most significant way of communication between normal people and hearing and speech impaired people without the need of an interpreter. Every country has its own developed Sign Language. In India, this dialect is known as Indian Sign Language. This research work aims at developing an automatic recognition system for Indian Sign Language numerals (0–9). The database used for implementation is self-created and consists of 1000 images, 100 images per numeral sign. Shape descriptors, Scale Invariant Feature Transform (SIFT) and Histogram of Oriented Gradients (HOG) techniques are used for extracting desired features. Artificial Neural Networks (ANN) and Support Vector Machine (SVM) classifiers are used to classify the signs. This system achieves accuracy as high as 99%.
2016 2nd International Conference on Next Generation Computing Technologies (NGCT), 2016
An effective method for highlight generation from basketball sports videos is proposed in this pa... more An effective method for highlight generation from basketball sports videos is proposed in this paper. Video shot classification is done which classify video into three categories that is close-up view, far view and replay. For close-up and far view detection we have proposed the feature based on the percentage of edge pixels to classify the frame into Far or close-up view. A pair of logos determines a replay segment and hence we have proposed an algorithm that determine logo and identify replay segments. Framework for ball detection is also described which is based on calculating intensity difference of two consecutive frames and performing morphology operations.
International Journal of Advanced Computer Science and Applications
Surveillance systems are widely used in malls, colleges, schools, shopping centers, airports, etc... more Surveillance systems are widely used in malls, colleges, schools, shopping centers, airports, etc. This could be due to the increasing crime rate in daily life. It is a very tedious task to monitor and detect abnormal activities 24x7 from the surveillance system. So the detection of abnormal events from videos is a hugely demanding area of research. In this paper, the proposed framework is used for deep learning concepts. Here SlowFast Resnet50 has been used to extract and process the features. After that, the deep neural network has been applied to generate a class using the Softmax function. The proposed framework has been applied to the UCF-Crime dataset using Graphics Processing Unit (GPU). It includes 1900 videos with 13 classes. Our proposed algorithm is evaluated by accuracy. Our proposed algorithm works better than the existing algorithm. It achieves 47.8% more accuracy than state of art method and also achieves good accuracy compared to other approaches used for detecting abnormal activity on the UCF-Crime dataset.
Digital image watermarking is one such technology that has been developed to protect digital imag... more Digital image watermarking is one such technology that has been developed to protect digital images from illegal manipulations. Watermarking methods are classified into two main classes: spatial and frequency based approaches. All methods are aim at satisfying two basic requirements: watermark perceptual visibility and watermark robustness against to attacks. In particular, digital image watermarking algorithms which are based on the discrete wavelet transform have been widely recognized to be more prevalent than others. This is due to the wavelets' excellent spatial localization, frequency spread, and multi-resolution characteristics, which are similar to the theoretical models of the human visual system. In This paper introduces an algorithm of digital watermarking based on Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) & applied various attacks and measure PSNR for Imperceptibility and NC values for Robustness.
Uploads
Papers by Mahasweta Joshi