Developing Methods to Increase Object Detection Success with Classical and Deep Learning Techniques

The aim of the project is to increase the overall success of the systems by improving the quality of image data used as input in artificial intelligence applications and to investigate and apply new generation image enhancement methods. In this context, all methods used in the literature to increase object detection success will be investigated and applied on a specified scenario. The developed methods are expected to have low hardware requirements and operate at high speed. Within the scope of the project, it is aimed to automatically determine the quality of the data set to be used in object detection, improve the quality of low quality data (Image Processing, Super Resolution, Image Restoration, etc.), eliminate data that does not meet the criteria, automatically determine the number of data sufficient for the object detection task, and complete the missing data with data augmentation methods or synthetic data generation. At the end of the project, it is expected to improve the quality of a low quality data set and to see a significant improvement in the object detection results obtained on the data set on the basis of numerical metrics such as average mean sensitivity.