Applied soft computing technologies

More about the book

SVMs have become highly effective pattern recognition methods, demonstrating superior performance in applications like text categorization, face detection, content-based image retrieval, and learning image similarity. The focus here is on detecting suspicious bags in security scenarios. Traditionally, bomb personnel resort to detonating suspicious bags, which can be catastrophic if the bags contain chemical, biological, or radiological materials. This method also destroys crucial evidence such as fingerprints and explosive signatures needed for forensic analysis. Utilizing telerobotics for extraction of bag contents mitigates these risks. In a telerobotic system, automating bag classification is beneficial, especially when combined with robotic tactics like shaking out contents. In cases where autonomous capabilities falter, human intervention may be necessary to identify bag types. This discussion emphasizes the autonomous operation of the telerobotic system, where SVMs classify bags to identify initial manipulator grasp points. Proper grasping requires the gripper to slide under the bag, but slippage can lead to failure. Therefore, understanding the bag type and the location of its opening is critical before assessing grasp points.