Semantic segmentation requires a detailed labeling of image pixels by object category. Information derived from local image patches is necessary to describe the detailed shape of individual objects. However, this information is ambiguous and can result in noisy labels. Global inference of image content can instead capture the general semantic concepts present. We advocate that holistic inference of image concepts provides valuable information for detailed pixel labeling. We propose a generic framework to leverage holistic information in the form of a LabelBank for pixel-level segmentation. We show the ability of our framework to improve semantic segmentation performance in a variety of settings. We learn models for extracting a holistic LabelBank from visual cues, attributes, and/or textual descriptions. We demonstrate improvements in semantic segmentation accuracy on standard datasets across a range of state-of-the-art segmentation architectures and holistic inference approaches.
We study large-scale kernel methods for acoustic modeling in speech recognition and compare their performance to deep neural networks (DNNs). We perform experiments on four speech recognition datasets, including the TIMIT and Broadcast News benchmark tasks, and compare these two types of models on frame-level performance metrics (accuracy, cross-entropy), as well as on recognition metrics (word/character error rate). In order to scale kernel methods to these large datasets, we use the random Fourier feature method of Rahimi and Recht (2007). We propose two novel techniques for improving the performance of kernel acoustic models. First, in order to reduce the number of random features required by kernel models, we propose a simple but effective method for feature selection. The method is able to explore a large number of non-linear features while maintaining a compact model more efficiently than existing approaches. Second, we present a number of frame-level metrics which correlate very strongly with recognition performance when computed on the heldout set; we take advantage of these correlations by monitoring these metrics during training in order to decide when to stop learning. This technique can noticeably improve the recognition performance of both DNN and kernel models, while narrowing the gap between them. Additionally, we show that the linear bottleneck method of Sainath et al. (2013) improves the performance of our kernel models significantly, in addition to speeding up training and making the models more compact. Together, these three methods dramatically improve the performance of kernel acoustic models, making their performance comparable to DNNs on the tasks we explored.
Zero-shot learning (ZSL) methods have been studied in the unrealistic setting where test data are assumed to come from unseen classes only. In this paper, we advocate studying the problem of generalized zero-shot learning (GZSL) where the test data's class memberships are unconstrained. We show empirically that naively using the classifiers constructed by ZSL approaches does not perform well in the generalized setting. Motivated by this, we propose a simple but effective calibration method that can be used to balance two conflicting forces: recognizing data from seen classes versus those from unseen ones. We develop a performance metric to characterize such a trade-off and examine the utility of this metric in evaluating various ZSL approaches. Our analysis further shows that there is a large gap between the performance of existing approaches and an upper bound established via idealized semantic embeddings, suggesting that improving class semantic embeddings is vital to GZSL.
We introduce a new multi-modal task for computer systems, posed as a combined vision-language comprehension challenge: identifying the most suitable text describing a scene, given several similar options. Accomplishing the task entails demonstrating comprehension beyond just recognizing "keywords" (or key-phrases) and their corresponding visual concepts. Instead, it requires an alignment between the representations of the two modalities that achieves a visually-grounded "understanding" of various linguistic elements and their dependencies. This new task also admits an easy-to-compute and well-studied metric: the accuracy in detecting the true target among the decoys. The paper makes several contributions: an effective and extensible mechanism for generating decoys from (human-created) image captions; an instance of applying this mechanism, yielding a large-scale machine comprehension dataset (based on the COCO images and captions) that we make publicly available; human evaluation results on this dataset, informing a performance upper-bound; and several baseline and competitive learning approaches that illustrate the utility of the proposed task and dataset in advancing both image and language comprehension. We also show that, in a multi-task learning setting, the performance on the proposed task is positively correlated with the end-to-end task of image captioning.
Semantic segmentation requires a detailed labeling of image pixels by object category. Information derived from local image patches is necessary to describe the detailed shape of individual objects. However, this information is ambiguous and can result in noisy labels. Global inference of image content can instead capture the general semantic concepts present. We advocate that high-recall holistic inference of image concepts provides valuable information for detailed pixel labeling. We build a two-stream neural network architecture that facilitates information flow from holistic information to local pixels, while keeping common image features shared among the low-level layers of both the holistic analysis and segmentation branches. We empirically evaluate our network on four standard semantic segmentation datasets. Our network obtains state-of-the-art performance on PASCAL-Context and NYUDv2, and ablation studies verify its effectiveness on ADE20K and SIFT-Flow.
Attention mechanisms have recently been introduced in deep learning for various tasks in natural language processing and computer vision. But despite their popularity, the "correctness" of the implicitly-learned attention maps has only been assessed qualitatively by visualization of several examples. In this paper we focus on evaluating and improving the correctness of attention in neural image captioning models. Specifically, we propose a quantitative evaluation metric for the consistency between the generated attention maps and human annotations, using recently released datasets with alignment between regions in images and entities in captions. We then propose novel models with different levels of explicit supervision for learning attention maps during training. The supervision can be strong when alignment between regions and caption entities are available, or weak when only object segments and categories are provided. We show on the popular Flickr30k and COCO datasets that introducing supervision of attention maps during training solidly improves both attention correctness and caption quality, showing the promise of making machine perception more human-like.
We propose a novel supervised learning technique for summarizing videos by automatically selecting keyframes or key subshots. Casting the problem as a structured prediction problem on sequential data, our main idea is to use Long Short-Term Memory (LSTM), a special type of recurrent neural networks to model the variable-range dependencies entailed in the task of video summarization. Our learning models attain the state-of-the-art results on two benchmark video datasets. Detailed analysis justifies the design of the models. In particular, we show that it is crucial to take into consideration the sequential structures in videos and model them. Besides advances in modeling techniques, we introduce techniques to address the need of a large number of annotated data for training complex learning models. There, our main idea is to exploit the existence of auxiliary annotated video datasets, albeit heterogeneous in visual styles and contents. Specifically, we show domain adaptation techniques can improve summarization by reducing the discrepancies in statistical properties across those datasets.
Given semantic descriptions of object classes, zero-shot learning aims to accurately recognize objects of the unseen classes, from which no examples are available at the training stage, by associating them to the seen classes, from which labeled examples are provided. We propose to tackle this problem from the perspective of manifold learning. Our main idea is to align the semantic space that is derived from external information to the model space that concerns itself with recognizing visual features. To this end, we introduce a set of "phantom" object classes whose coordinates live in both the semantic space and the model space. Serving as bases in a dictionary, they can be optimized from labeled data such that the synthesized real object classifiers achieve optimal discriminative performance. We demonstrate superior accuracy of our approach over the state of the art on four benchmark datasets for zero-shot learning, including the full ImageNet Fall 2011 dataset with more than 20,000 unseen classes.
Video summarization has unprecedented importance to help us digest, browse, and search today's ever-growing video collections. We propose a novel subset selection technique that leverages supervision in the form of human-created summaries to perform automatic keyframe-based video summarization. The main idea is to nonparametrically transfer summary structures from annotated videos to unseen test videos. We show how to extend our method to exploit semantic side information about the video's category/genre to guide the transfer process by those training videos semantically consistent with the test input. We also show how to generalize our method to subshot-based summarization, which not only reduces computational costs but also provides more flexible ways of defining visual similarity across subshots spanning several frames. We conduct extensive evaluation on several benchmarks and demonstrate promising results, outperforming existing methods in several settings.
We study large-scale kernel methods for acoustic modeling and compare to DNNs on performance metrics related to both acoustic modeling and recognition. Measuring perplexity and frame-level classification accuracy, kernel-based acoustic models are as effective as their DNN counterparts. However, on token-error-rates DNN models can be significantly better. We have discovered that this might be attributed to DNN's unique strength in reducing both the perplexity and the entropy of the predicted posterior probabilities. Motivated by our findings, we propose a new technique, entropy regularized perplexity, for model selection. This technique can noticeably improve the recognition performance of both types of models, and reduces the gap between them. While effective on Broadcast News, this technique could be also applicable to other tasks.