Character recognition is right now achieving attention of most of the researchers because of its huge applications in various sectors like human-robot interaction, data entry for business documents, etc. Recognition of characters is tricky task, but Deep learning approach can be adequately used as a solution for various such problems. Discrepancy in writing style makes handwritten character recognition one of the most burdensome work. There are plenty ways to write a single letter or a digit which automatically increases the size of the dataset to be used. The goal of this work is to integrate machine learning techniques to improve the character recognition process.

Since the past few years, character recognition is gaining attention of most of the researchers because of its various applications in different sectors like automatic recognition of vehicle number plate, data entry for business documents for example cheque, bank statement, passport, invoice and receipt, in airports for passport recognition and information extraction. There are many variations in writing styles from individual to individual which makes character recognition challenging. Deep learning techniques can be used to find solutions for recognition of characters using different algorithms. A single letter or digit can be written in many ways and styles which increases the size of the dataset to be utilized.The goal of this work is to join AI methods to enhance and improve the character recognition process.

Medical Image Processing may be a complicated and challenging area in recent times. Processing of MRI image is one of the parts of this discipline. This paper proposes a method for green detection of a brain tumour in MRI brain images. The methodology consists of the subsequent steps: Pre-processing by using median filter & segmentation of the image is performed by thresholding. This technique is then accompanied via the further application of morphological operations. As part of the defined operations, the outputs are recorded. The recorded outputs are in a position to distinguish between the affected and the non-affected elements of the brain tumour.

Brain tumor detection is one of the most complex problems due to its anatomical structure. Medical segmentation has always been the challenging part in curing 6x brain tumors. In such cases, deep learning algorithms have made it easier to perform segmentation such as Convolutional Neural Networks (CNN): UNET, Bit plane method. Using these methods of computer visions we have increased the rate of successfully detecting the tumor. Extensive use of biomedical image segmentation has resulted in acquiring accurate results which has given high rates of curing the tumor. In this paper, we propose to combine different techniques like 3D MRI, UNET architecture and Bit plane method to perform segmentation of Enhancing Tumor, Tumor Core and Whole tumor which are the subregions of Gliomas.

Low visibility is the drawback in road transport because of particles, smoke, dust, fog, rain, and wetness that are suspended within the air. We tend to refer to as haze that reduces visibility. A boundary Constraint and discourse Regularization de-hazing algorithmic rule is planned within which we tend to use one frame image for enhancing the foggy image. Aftermanypieces of research, a tool has been developed Driver Assistance Systems (DAS) supported in integrated systems of the vehicle. Thisproject aims to assist drivers in hazy driving conditions.