A5044484189- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Autonomous Vehicle Technology and Safety
- Video Surveillance and Tracking Methods
- Vehicle License Plate Recognition
- Currency Recognition and Detection
- Bone Tumor Diagnosis and Treatments
- Satellite Image Processing and Photogrammetry
- Sarcoma Diagnosis and Treatment
- Medical Imaging and Analysis
- Management of metastatic bone disease
- Traffic Prediction and Management Techniques
- Video Analysis and Summarization
- Industrial Vision Systems and Defect Detection
- Image and Object Detection Techniques
- Dental Radiography and Imaging
- Medical Imaging and Pathology Studies
- Traffic control and management
- Hand Gesture Recognition Systems
- Image Processing and 3D Reconstruction
- Smart Parking Systems Research
- Oral and Maxillofacial Pathology
- Remote Sensing and LiDAR Applications
- Infrared Thermography in Medicine
Institute of Infrastructure Technology Research and Management
2018-2024
Indian Institute of Technology Bombay
2011-2016
In the proposed work, initially, YOLO algorithm is used to extract and classify objects in a frame. sequence of frames, due various reasons confidence measure suddenly drops. This changes class an object consecutive frames which affects tracking counting process severely. To overcome this limitation algorithm, it modified enable track same efficiently frames. will turn increase accuracy. work drastic change scores are identified by particular These outliers detected removed using RANSAC...
An efficient product packing system through a simplified automatic counting and collecting process is critical aspect of the manufacturing industry. In present paper, we propose real-time tool for simultaneous collection objects into different bins using machine vision. The employs minimum distance classifier object detection, Counting on conveyor belt done by tracking Euclidian distances between centroids in successive frames. Furthermore, required number carried out two steps. first step,...
Pathological conditions in the knee joint can be caused by vibrations emitted while bending or extending leg. To detect this to know inner condition of like softening, coarseness, breakage it's state lubrication articular cartilage surface, vibroarthrographic (VAG) signal is useful. The VAG used because fluctuating and nonlinear property which helpful extract joint. In present work on signals using TQWT decomposition. are partitioned into sub-band distinct frequencies. There different...
In remote sensing linear scanners are widely used for acquiring stereo images as they provide data with very high resolution. A trinocular view increases the accuracy of 3D modeling it gives an additional to compensate occlusions. However in triplet mode by is challenged processing steps which lead a new research area society. various earth and planetary missions such ALOS PRISM, Chang'e-1, Pléiades Chandrayaan-1 sensor was capture mode. The pushbroom scanner 3 array sensors looking three...
Intersection study is one of the crucial surveys for advanced traffic management, especially at unsignalized points. In this study, traditional method vehicle detection and classification initially implemented to classify volume intersections. To enhance accuracy object counting, a deep learning approach counting designed using YOLO (You Only Look Once) algorithm modified algorithm. This includes variation statistical best outlier detection. The aims improve efficiency intersections, which...
Introduction Panoramic dental images gives an in-depth understanding of the tooth structure, both lower and upper jaws, surrounding structures throughout cavity in our mouth.The provided have significance for diagnostics since they aid detection array disorders, including oral cancer. We propose a novel approach to automatic joint teeth segmentation using pioneer Mask Recurrent Convolutional Neural Network (MRCNN) model image segmentation. Material Methods In this study, sequence residual...
In the proposed method, an innovative image processing technique for vehicle tracking at a roundabout is described. Background subtraction applied to get objects (vehicles) in foreground. The are thus obtained and tracked using morphological operations object properties. video stream, established by tracing center of target object. present case,four different directions incoming traffic considered four classes defined. Implementation above mentioned method achieved promising result accuracy...
The aim of this research is to automatize handwritten attendance data entries in digital form with a minimal human intervention using Handwritten Character Recognition (HCR). Here, all the preprocessing techniques (Image Registration, Cropping, Erosion. Dilation, Noise removal, Line Segmentation), and classification technique are implemented MATLAB platform. Various types for recognition digits characters have been different applications. sheet classified Structural Similarity Measurement...
Abstract. In this paper, we propose a new scan line optimization method for matching the triplet of images. present triplets are initially matched using an area based local method. The cost is stored in structure called as Disparity Space Image (DSI). Using global minimum initial disparity generated. Next minima considered potential matches where gives erroneous results. These used disparity. As scanned optimization, it use popularly resampled experiment performed Terrain Mapping Camera...
In this paper a method is proposed to match corners for stereo correspondence. Corners convey lot of information, using which they can be matched. The investigates the importance as potential points accurate disparity map generation. Determining correspondence at in image pair known well enough generate sparse map, same time are very unstable slight shift view point changes background corner. Therefore area and feature based methods limited corner with lower accuracy. matching algorithm...
In this paper a novel approach of stereoscopic depth inpainting is used to model lunar craters. the case mission, disparity map generation using images limited by shadows inside impact Impact craters are most important features on surface but cannot be modeled stereo cue alone. Taking into consideration importance crater modeling and limitations cue, we proposing method as an additional crater.