Attention Awareness Multiple Instance Neural Network

Multiple instance learning is qualified for many pattern recognition tasks with weakly annotated data. The combination of artificial neural network and multiple offers an end-to-end solution has been widely utilized. However, challenges remain in two-folds. Firstly, current MIL pooling operators are usually pre-defined lack flexibility to mine key instances. Secondly, solutions, the bag-level representation can be inaccurate or inaccessible. To this end, we propose attention awareness framework paper. It consists instance-level classifier, a trainable operator based on spatial classification layer. Exhaustive experiments series demonstrate that our outperforms state-of-the-art methods validates effectiveness proposed operators.