Instance segmentation is a challenging task aiming at classifying and segmenting all object instances of specific classes. While two-stage box-based methods achieve top performances in the image domain, they cannot easily extend their superiority into video domain. This because usually deal with features or images cropped from detected bounding boxes without alignment, failing to capture pixel-...

Confounded information is an objective fact when using multi-instance learning (MIL) to classify bags of instances, which may be inherited by MIL embedding methods and lead questionable bag label prediction. To respond this problem, we propose the with deconfounded instance-level prediction algorithm. Unlike traditional embedding-based strategies, design a optimization goal maximize distinction...

When applying multi-instance learning (MIL) to make predictions for bags of instances, the prediction accuracy an instance often depends on not only itself but also its context in corresponding bag. From viewpoint causal inference, such bag contextual prior works as a confounder and may result model robustness interpretability issues. Focusing this problem, we propose novel interventional (IMIL...

Multiple Instance Learning (MIL) is concerned with learning from sets (bags) of feature vectors (instances), where the individual instance labels are ambiguous. In MIL it is often assumed that positive bags contain at least one instance from a so-called concept in instance space, whereas negative bags only contain negative instances. The classes in a MIL problem are therefore not treated in the...

Instance search is an interesting task as well a challenging issue due to the lack of effective feature representation. In this paper, instance level representation built upon fully convolutional instance-aware segmentation proposed. The ROI-pooled from segmented region. So that instances in various sizes and layouts are represented by deep features uniform length. This further enhanced use def...

We propose a multi-cue based approach for recognizing human actions in still images, where relevant object regions are discovered and utilized in a weakly supervised manner. Our approach does not require any explicitly trained object detector or part/attribute annotation. Instead, a multiple instance learning approach is used over sets of object hypotheses in order to represent objects relevant...

We present a new multiple-instance (MI) learning technique (EMDD) that combines EM with the diverse density (DD) algorithm. EM-DD is a general-purpose MI algorithm that can be applied with boolean or real-value labels and makes real-value predictions. On the boolean Musk benchmarks, the EM-DD algorithm without any tuning significantly outperforms all previous algorithms. EM-DD is relatively ins...