| Popis: |
In the ever-evolving world of computer vision, image recognition is critical task including using image processing to solve real-world problems like reducing human involvement in the art of driving. We face many challenges in completing this mission, including object detection and segmentation. By integrating Keras and TensorFlow with instance segmentation and binary masks, the challenges are effectively overcome by the method proposed in this chapter. The technique instantaneous segmentation is adopted for separating and detecting each individual object of interest in an image based on their pixel characteristics. The Mask RCNN model outperforms the current CNN models in terms of object detection accuracy and performance. Also, the objects of interest are classified using regional proposal networks (RPN) instead of selective search algorithm by CNN. The instance segmentation system is conceptually clear, versatile, and general. The method successfully finds items in acquired input and also produces good results by masking for each case. |
