Video-Based Sign Language Recognition Using CNN for Frame Feature Extraction and a Hybrid Temporal Model with Self-Attention and LSTM Sakshi Kharat, Bhakti Palkar, Mansi Kambli Advances in Science and Technology, 2026 The main form of communication for hearing-impaired individuals is sign language, although non-signers face challenges in understanding it. To bridge this gap, we propose a hybrid model designed to recognize sign gestures using the LSA64 dataset. Our approach employs a custom 3D-CNN to extract spatiotemporal features, integrated with a hybrid sequence model to capture temporal dynamics. This sequence model combines Temporal Convolutional Layer, Self-Attention Mechanism for frame-to-frame relationships, and a LSTM network for modeling long-term dependencies. We compared our model against six-pretrained feature extractors, all using the same sequence modeling architecture. The experimental results indicate that our proposed model surpasses all the baselines with an accuracy of 90.31% thereby demonstrating its effectiveness in recognizing sign gestures.
Sugarcane Categorisation using Deep Learning Models on Sentinel Dataset Mansi Kambli, Bhakti Palkar Technologies for Energy Agriculture and Healthcare, 2025 This paper’s primary goal is to classify sugarcane crops using deep learning models using Sentinel imaging data. The sugarcane crop is a cash crop used in the production of ethanol and sugar. The classification of sugarcane is crucial for agricultural oversight and management. The conventional crop classification techniques that rely on limited ground-based data collection or manual examination take a lot of time and are usually unreliable. Consequently, an automated and effective approach is proposed, that requires the utilisation of imagery from satellites data and the deep learning techniques. The Convolutional neural network, ResNet and Inception-ResNet are the models applied for sugarcane classification using multispectral satellite data. The categorisation helps the farmers in timely decisions and management of the crop for good yield. The remote sensing through Sentinel data warns the farmers for pests and early detection of disease as categorisation of sugarcane is done prior. Also helps in monitoring the crop for better yield.
WGAN-CNN for dense and sparse vegetation categorisation of remotely sensed sugarcane crop Mansi Manoj Kambli, Bhakti Palkar 2025 IEEE International Conference on Next Gen Technologies of Artificial Intelligence and Geoscience Remote Sensing Earthsense 2025, 2025 Sugarcane is India’s second largest crop and requires extensive monitoring for optimal health and productivity due to its 12-15 month growth duration. Data sets for the sugarcane crop are also not available publicly. For this, a new approach is applied in this article to increase the dataset of dense and sparse vegetation of the Napane region. The Normalized Difference Vegetation Index (NDVI) of the Napane region was used using Sentinel-2 maps with ground truth data to identify dense and sparse vegetation in the spectral bands. The proposed Wasserstein Generative Adversarial Network (WGAN) with batch normalization generated 10000 more images for dense and sparse, respectively, by stabilizing the training process and reducing gradient vanishing. After applying the CNN model to the generated images, a 95.63 % accuracy rate is achieved. The categorization of dense and sparse vegetation helps farmers monitor the growth of sugarcane crops that leads to sustainable agriculture.
Categorisation of Vegetation Using Machine Learning and Remote Sensing Methods Mansi Kambli Journal of Image and Graphics United Kingdom, 2025 —Farming is the main source of income, so crop development must be continuously watched and given the highest attention possible for the farmers. Precision agriculture is a farming management approach that uses multispectral satellite data to monitor, measure, and adjust to temporal and geographical variability in order to improve the sustainability of agricultural output. Sugarcane is a cash crop and is used in this study as researchers are focusing more on success in sugarcane development. Detecting dense and sparse vegetation for the individual plots of the farmers helps to understand that the plot area has favorable soil and water conditions for sugarcane growth. The sparse vegetation indicates that the area has slopes and water is not retained creating problems in the growth of the sugarcane. The Remote Sensing spectral bands are used and the vegetation indices like Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Ratio Vegetation Index (RVI) are used in the sugarcane canopy. EVI works for dense canopy and RVI for sparse vegetation as shown by the research done in this paper. The Machine Learning (ML) further also helps to detect the sparse and dense vegetation and the accuracy of all the classifiers is compared for the same. The survey on different Machine Learning techniques applied to remotely sensed data of sugarcane crops is done in this research work. This study aims to monitor sugarcane crop health by detecting sparse and dense vegetation using vegetation indices, and it evaluates the performance of different ML classifiers for precision agriculture. The same plot of the farmer can be monitored each month to find the change detection and further, the cause of sparse vegetation in the particular plot can be diagnosed with the help of enhanced vegetation indices in future work. To locate healthy vegetation, RS Sensing uses the Normalised Difference Vegetation Index (NDVI) and it gets saturated at grand growth stages so the novel method is enhance vegetation indices and ratio vegetation index which can be used to monitor at grand growth stages along with ML models as shown in this research work.
Sortitude - AI Email Sorter 16th International Conference on Advances in Computing Control and Telecommunication Technologies Act 2025, 2025
DoodSearch - OpenCV with Image Recognition Harshit Jain, Kashish Harisinghani, Shubh Gangar, Mansi Kambli 2022 IEEE Conference on Interdisciplinary Approaches in Technology and Management for Social Innovation Iatmsi 2022, 2022 Hand recognition using OpenCV is a burgeoning field with several applications, including virtual environment control, sign language translation, robot remote control, and musical composition. One such application is gesture recognition, which allows users to practically create anything on the screen by simply using their hands. Another exciting function that is prevalent is searching by image. The concept behind searching by image is to take a picture as an input and generate metadata from it that may be used later. In this paper, integration of both the aforementioned methodologies is done to produce an application in which a user may draw any image using numerous hand motions that are captured by the user's webcam. This image is then processed by an image recognition algorithm, which generates metadata that may be used to search for the entity depicted in the image.
Lossy Image Compression-A Comparison Between Wavelet Transform, Principal Component Analysis, K-Means and Autoencoders Ankit Thakker, Nikhil Namboodiri, Ritik Mody, Riya Tasgaonkar, Mansi Kambli 5th IEEE International Conference on Advances in Science and Technology Icast 2022, 2022 For many years, lossless image compression has been a promising topic of study. Various techniques have been created over time to obtain an approximation of the reduced data size. While discrete wavelet transform (DWT) and discrete cosine transform (DCT) have historically been employed for the purpose of compressing images, various machine learning methods and deep learning networks are now being offered. In this research, we conduct a comparative analysis of conventional and contemporary lossy image compression techniques on the Kodak Dataset, including Autoencoders, Principal Component Analysis (PCA), K-Means, and Discrete Wavelet Transform (DWT). The metrics used for the evaluation of the proposed study are Mean Squared Error (MSE), Peak Signal to Noise Ratio (PSNR), Compression Ratio (CR), and Structural Similarity Index (SSIM).
Automatic Question Classifier Neha Patil, Ojas Kulkarni, Vineeta Bhujle, Aniket Joshi, Khushi Khanchandani, Mansi Kambli Proceedings of 4th International Conference on Cybernetics Cognition and Machine Learning Applications Icccmla 2022, 2022
Design and Implementation of Hybrid BCI based Wheelchair Poonam Chawda, Arvind Sridhar, Arvind Gyandatt Mishra, Heeral Dedhia, Mansi Kambli, Sushma Kadge Proceedings 2nd International Conference on Smart Electronics and Communication Icosec 2021, 2021
Internet of things based intelligent parking system facilitated by third party payment platform 10th International Conference on Advances in Computing Control and Telecommunication Technologies Act 2019, 2019
IOT Based Solution to Reduce Vehicular Traffic Congestion 12th Indiacom 5th International Conference on Computing for Sustainable Global Development Indiacom 2018, 2018
An IoT based approach for Urban Water Management System 12th Indiacom 5th International Conference on Computing for Sustainable Global Development Indiacom 2018, 2018
Smart Home Automation System 12th Indiacom 5th International Conference on Computing for Sustainable Global Development Indiacom 2018, 2018
