For learned image representations, basic autoencoders often produce blurry results. Reconstruction quality can be improved by incorporating additional penalties such as adversarial (GAN) and perceptual losses. Arguably, these approaches lack a principled interpretation. Concurrently, in generative settings diffusion has demonstrated a remarkable ability to create crisp, high quality results and has solid theoretical underpinnings (from variational inference to direct study as the Fisher Divergence). Our work combines autoencoder representation learning with diffusion and is, to our knowledge, the first to demonstrate jointly learning a continuous encoder and decoder under a diffusion-based loss and showing that it can lead to higher compression and better generation. We demonstrate that this approach yields better reconstruction quality as compared to GAN-based autoencoders while being easier to tune. We also show that the resulting representation is easier to model with a latent diffusion model as compared to the representation obtained from a state-of-the-art GAN-based loss. Since our decoder is stochastic, it can generate details not encoded in the otherwise deterministic latent representation; we therefore name our approach "Sample what you can't compress", or SWYCC for short.

Ensembling has proven to be a powerful technique for boosting model performance, uncertainty estimation, and robustness in supervised learning. Advances in self-supervised learning (SSL) enable leveraging large unlabeled corpora for state-of-the-art few-shot and supervised learning performance. In this paper, we explore how ensemble methods can improve recent SSL techniques by developing a framework that permits data-dependent weighted cross-entropy losses. We refrain from ensembling the representation backbone; this choice yields an efficient ensemble method that incurs a small training cost and requires no architectural changes or computational overhead to downstream evaluation. The effectiveness of our method is demonstrated with two state-of-the-art SSL methods, DINO (Caron et al., 2021) and MSN (Assran et al., 2022). Our method outperforms both in multiple evaluation metrics on ImageNet-1K, particularly in the few-shot setting. We explore several weighting schemes and find that those which increase the diversity of ensemble heads lead to better downstream evaluation results. Thorough experiments yield improved prior art baselines which our method still surpasses; e.g., our overall improvement with MSN ViT-B/16 is 3.9 p.p. for 1-shot learning.

While batch normalization has been successful in speeding up the training of neural networks, it is not well understood. We cast batch normalization as an approximation of the limiting case where the entire dataset is normalized jointly, and explore other ways to approximate the gradient from this limiting case. We demonstrate an approximation that removes the need to keep more than one example in memory at any given time, at the cost of a small factor increase in the training step computation, as well as a fully per-example training procedure, which removes the extra computation at the cost of a small drop in the final model accuracy. We further use our insights to improve batch renormalization for very small minibatches. Unlike previously proposed methods, our normalization does not change the function class of the inference model, and performs well in the absence of identity shortcuts.

Methods, systems, and media for summarizing a video with video thumbnails are provided. In some embodiments, the method comprises: receiving a plurality of video frames corresponding to the video and associated information associated with each of the plurality of video frames; extracting, for each of the plurality of video frames, a plurality of features; generating candidate clips that each includes at least a portion of the received video frames based on the extracted plurality of features and the associated information; calculating, for each candidate clip, a clip score based on the extracted plurality of features from the video frames associated with the candidate clip; calculating, between adjacent candidate clips, a transition score based at least in part on a comparison of video frame features between frames from the adjacent candidate clips; selecting a subset of the candidate clips based at least in part on the clip score and the transition score associated with each of the candidate clips; and automatically generating an animated video thumbnail corresponding to the video that includes a plurality of video frames selected from each of the subset of candidate clips.

um sistema de rede neural que inclui sub-redes, incluindo uma primeira sub-rede, incluindo varios primeiros modulos, cada primeiro modulo incluindo: uma camada convolucional de passagem que processa a entrada da sub-rede para gerar uma saida de passagem; uma pilha comum de camadas de rede neural que processa a entrada da sub-rede para a primeira sub-rede para gerar uma saida de agrupamento media; uma primeira pilha de camadas de rede neural convolucional que processa a entrada da sub-rede para gerar uma primeira saida de pilha; uma segunda pilha de camadas de rede neural convolucional que processa a entrada da sub-rede para gerar uma segunda saida de pilha; e uma camada de concatenacao que concatena as saidas geradas para gerar uma saida do primeiro modulo. ao incluir sub-redes de modulo em uma rede neural profunda, a rede neural profunda pode ter melhor desempenho nas tarefas de processamento de imagem e pode ser treinada de maneira mais rapida e eficiente, mantendo um desempenho aprimorado nas tarefas de processamento de imagem.

We present a full reference, perceptual image metric based on VGG-16, an artificial neural network trained on object classification. We fit the metric to a new database based on 140k unique images annotated with ground truth by human raters who received minimal instruction. The resulting metric shows competitive performance on TID 2013, a database widely used to assess image quality assessments methods. More interestingly, it shows strong responses to objects potentially carrying semantic relevance such as faces and text, which we demonstrate using a visualization technique and ablation experiments. In effect, the metric appears to model a higher influence of semantic context on judgments, which we observe particularly in untrained raters. As the vast majority of users of image processing systems are unfamiliar with Image Quality Assessment (IQA) tasks, these findings may have significant impact on real-world applications of perceptual metrics.

Batch Normalization is quite effective at accelerating and improving the training of deep models. However, its effectiveness diminishes when the training minibatches are small, or do not consist of independent samples. We hypothesize that this is due to the dependence of model layer inputs on all the examples in the minibatch, and different activations being produced between training and inference. We propose Batch Renormalization, a simple and effective extension to ensure that the training and inference models generate the same outputs that depend on individual examples rather than the entire minibatch. Models trained with Batch Renormalization perform substantially better than batchnorm when training with small or non-i.i.d. minibatches. At the same time, Batch Renormalization retains the benefits of batchnorm such as insensitivity to initialization and training efficiency.

Very deep convolutional networks have been central to the largest advances in image recognition performance in recent years. One example is the Inception architecture that has been shown to achieve very good performance at relatively low computational cost. Recently, the introduction of residual connections in conjunction with a more traditional architecture has yielded state-of-the-art performance in the 2015 ILSVRC challenge; its performance was similar to the latest generation Inception-v3 network. This raises the question: Are there any benefits to combining Inception architectures with residual connections? Here we give clear empirical evidence that training with residual connections accelerates the training of Inception networks significantly. There is also some evidence of residual Inception networks outperforming similarly expensive Inception networks without residual connections by a thin margin. We also present several new streamlined architectures for both residual and non-residual Inception networks. These variations improve the single-frame recognition performance on the ILSVRC 2012 classification task significantly. We further demonstrate how proper activation scaling stabilizes the training of very wide residual Inception networks. With an ensemble of three residual and one Inception-v4 networks, we achieve 3.08% top-5 error on the test set of the ImageNet classification (CLS) challenge.

We address the problem of distance metric learning (DML), defined as learning a distance consistent with a notion of semantic similarity. Traditionally, for this problem supervision is expressed in the form of sets of points that follow an ordinal relationship - an anchor point x is similar to a set of positive points Y, and dissimilar to a set of negative points Z, and a loss defined over these distances is minimized. While the specifics of the optimization differ, in this work we collectively call this type of supervision Triplets and all methods that follow this pattern Triplet-Based methods. These methods are challenging to optimize. A main issue is the need for finding informative triplets, which is usually achieved by a variety of tricks such as increasing the batch size, hard or semi-hard triplet mining, etc. Even with these tricks, the convergence rate of such methods is slow. In this paper we propose to optimize the triplet loss on a different space of triplets, consisting of an anchor data point and similar and dissimilar proxy points which are learned as well. These proxies approximate the original data points, so that a triplet loss over the proxies is a tight upper bound of the original loss. This proxy-based loss is empirically better behaved. As a result, the proxy-loss improves on state-of-art results for three standard zero-shot learning datasets, by up to 15% points, while converging three times as fast as other triplet-based losses.

Training Deep Neural Networks is complicated by the fact that the distribution of each layer’s inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it no toriously hard to train models with saturating nonlinearities. We refer to this phenomenon as internal covariate shift, and address the problem by normalizing layer inputs. Our method draws its strength from making normalization a part of the model architecture and performing the normalizationfor each training mini-batch . Batch Normalization allows us to use much higher learning rates and be less careful about initialization. It also acts as a regularizer, in some cases eliminating the need for Dropout. Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin. Using an ensemble of batchnormalized networks, we improve upon the best published result on ImageNet classification: reaching 4.9% top-5 validation error (and 4.8% test error), exceeding the accuracy of human raters.

A method and apparatus for performing sparse coding dictionary priming are disclosed. Sparse coding dictionary priming may include iteratively training a coding dictionary, which may include a plurality of codewords or bases. Iteratively training the coding dictionary may include identifying a sampling index cardinality, identifying a portion of a video stream, decomposing the portion of the video stream, and updating the codeword based on the portion of the video stream. Decomposing the portion of the video stream may include randomly identifying a set of codewords from the plurality of codewords wherein a cardinality of the set of codewords is the sampling index cardinality and wherein the sampling index cardinality is less a cardinality of the plurality of codewords, and determining a codeword having a maximum correlation with the portion of the video stream from the set of codewords.

Convolutional networks are at the core of most state of-the-art computer vision solutions for a wide variety of tasks. Since 2014 very deep convolutional networks started to become mainstream, yielding substantial gains in various benchmarks. Although increased model size and computational cost tend to translate to immediate quality gains for most tasks (as long as enough labeled data is provided for training), computational efficiency and low parameter count are still enabling factors for various use cases such as mobile vision and big-data scenarios. Here we are exploring ways to scale up networks in ways that aim at utilizing the added computation as efficiently as possible by suitably factorized convolutions and aggressive regularization. We benchmark our methods on the ILSVRC 2012 classification challenge validation set demonstrate substantial gains over the state of the art: 21:2% top-1 and 5:6% top-5 error for single frame evaluation using a network with a computational cost of 5 billion multiply-adds per inference and with using less than 25 million parameters. With an ensemble of 4 models and multi-crop evaluation, we report 3:5% top-5 error and 17:3% top-1 error on the validation set and 3:6% top-5 error on the official test set.

A video server receives an uploaded video and determines whether the video contains third-party content and which portions of the uploaded video match third-party content. The video server determines whether to degrade the matching portions and/or how (e.g., extent, type) to do so. The video server separates the matching portion from original portions in the uploaded video and generates a degraded version of the matching content by applying an effect such as compression, edge distortion, temporal distortion, noise addition, color distortion, or audio distortion. The video server combines the degraded portions with the original portions to output a degraded version of the uploaded video. The video server stores and/or distributes the degraded version of the uploaded video. The video server may offer the uploading user licensing terms with the content owner that the user may accept to reverse the degradation.