We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

We present a conceptually simple, flexible, and general framework for object instance segmentation. Our approach efficiently detects objects in an image while simultaneously generating a high-quality segmentation mask for each instance. The method, called Mask R-CNN, extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition. Mask R-CNN is simple to train and adds only a small overhead to Faster R-CNN, running at 5 fps. Moreover, Mask R-CNN is easy to generalize to other tasks, e.g., allowing us to estimate human poses in the same framework. We show top results in all three tracks of the COCO suite of challenges, including instance segmentation, bounding-box object detection, and person keypoint detection. Without tricks, Mask R-CNN outperforms all existing, single-model entries on every task, including the COCO 2016 challenge winners. We hope our simple and effective approach will serve as a solid baseline and help ease future research in instance-level recognition. Code will be made available.

Mask R-CNN

While the Transformer architecture has become the de-facto standard for natural language processing tasks, its applications to computer vision remain limited. In vision, attention is either applied in conjunction with convolutional networks, or used to replace certain components of convolutional networks while keeping their overall structure in place. We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks. When pre-trained on large amounts of data and transferred to multiple mid-sized or small image recognition benchmarks (ImageNet, CIFAR-100, VTAB, etc.), Vision Transformer (ViT) attains excellent results compared to state-of-the-art convolutional networks while requiring substantially fewer computational resources to train.

An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

We present a conceptually simple, flexible, and general framework for object instance segmentation. Our approach efficiently detects objects in an image while simultaneously generating a high-quality segmentation mask for each instance. The method, called Mask R-CNN, extends Faster R-CNN by adding a branch for predicting an object mask in parallel with the existing branch for bounding box recognition. Mask R-CNN is simple to train and adds only a small overhead to Faster R-CNN, running at 5 fps. Moreover, Mask R-CNN is easy to generalize to other tasks, e.g., allowing us to estimate human poses in the same framework. We show top results in all three tracks of the COCO suite of challenges, including instance segmentation, bounding-box object detection, and person keypoint detection. Without bells and whistles, Mask R-CNN outperforms all existing, single-model entries on every task, including the COCO 2016 challenge winners. We hope our simple and effective approach will serve as a solid baseline and help ease future research in instance-level recognition. Code has been made available at: this https URL

Despite significant recent advances in the field of face recognition [10, 14, 15, 17], implementing face verification and recognition efficiently at scale presents serious challenges to current approaches. In this paper we present a system, called FaceNet, that directly learns a mapping from face images to a compact Euclidean space where distances directly correspond to a measure of face similarity. Once this space has been produced, tasks such as face recognition, verification and clustering can be easily implemented using standard techniques with FaceNet embeddings as feature vectors.

/pdf/facenet-a-unified-embedding-for-face-recognition-and-135p2gog88.pdf

FaceNet: A unified embedding for face recognition and clustering

Photo collage is an effective representation for photo summarization and visualization. However, its practical usage is limited by its high computational cost. In this paper, we present efficient optimization techniques based on a novel formulation on markov random fields and several insights unexploited by previous approaches. Our method achieves up to hundreds-fold performance improvement, and user/application specific constraints can be easily integrated. We present new applications including interactive collage refinement and dynamic collage for photo browsing. Preliminary evaluation and user study indicates that our approach extends the usability of photo collage.

/pdf/efficient-optimization-of-photo-collage-5bt7mvgawi.pdf

Efficient Optimization of Photo Collage

Techniques for unsupervised object class discovery via bottom-up multiple class learning are described. These techniques may include receiving multiple images containing one or more object classes. The multiple images may be analyzed to extract top saliency instances and least saliency instances. These saliency instances may be clustered to generate and/or update statistical models. The statistical models may be used to discover the one or more object classes. In some instances, the statistical models may be used to discover object classes of novel images.

Unsupervised object class discovery via bottom up multiple class learning

There has been significant progress on pose estimation and increasing interests on pose tracking in recent years. At the same time, the overall algorithm and system complexity increases as well, making the algorithm analysis and comparison more difficult. This work provides simple and effective baseline methods. They are helpful for inspiring and evaluating new ideas for the field. State-of-the-art results are achieved on challenging benchmarks. The code will be available at this https URL.

Simple Baselines for Human Pose Estimation and Tracking

Described is a technology in which a collage of digital photographs is dynamically computed and rendered so as to vary the photographs that are visible in the collage over time. A dynamic collage mechanism coupled to a source of photographs computes a collage for visible output, and dynamically updates the collage on a scheduled basis by adding different photograph(s) in place of other photograph(s). Each arrangement of the photographs in each updated collage is computed from a previous collage. Also described is layout optimization in which the photographs in the updated collage are translated, rotated and/or layered so as to cover a maximum amount of the overall area of the collage.

Dynamic Collage for Visualizing Large Photograph Collections

This paper introduces a new fundamental characteristic, i.e., the dynamic range, from real-world metric tools to deep visual recognition. In metrology, the dynamic range is a basic quality of a metric tool, indicating its flexibility to accommodate various scales. Larger dynamic range offers higher flexibility. In visual recognition, the multiple scale problem also exist. Different visual concepts may have different semantic scales. For example, "Animal" and "Plants" have a large semantic scale while "Elk" has a much smaller one. Under a small semantic scale, two different elks may look quite different to each other . However, under a large semantic scale (e.g., animals and plants), these two elks should be measured as being similar.Introducing the dynamic range to deep metric learning, we get a novel computer vision task, i.e., the Dynamic Metric Learning. It aims to learn a scalable metric space to accommodate visual concepts across multiple semantic scales. Based on three types of images, i.e., vehicle, animal and online products, we construct three datasets for Dynamic Metric Learning. We benchmark these datasets with popular deep metric learning methods and find Dynamic Metric Learning to be very challenging. The major difficulty lies in a conflict between different scales: the discriminative ability under a small scale usually compromises the discriminative ability under a large one, and vice versa. As a minor contribution, we propose Cross-Scale Learning (CSL) to alleviate such conflict. We show that CSL consistently improves the baseline on all the three datasets. The datasets and the code will be publicly available at https://github.com/SupetZYK/DynamicMetricLearning.

/pdf/dynamic-metric-learning-towards-a-scalable-metric-space-to-1qgbrs7pzc.pdf

Yichen Wei

Papers

Efficient Optimization of Photo Collage

Unsupervised object class discovery via bottom up multiple class learning

Simple Baselines for Human Pose Estimation and Tracking

Dynamic Collage for Visualizing Large Photograph Collections

Dynamic Metric Learning: Towards a Scalable Metric Space to Accommodate Multiple Semantic Scales