Single image inverse problem is a notoriously challenging ill-posed problem that aims to restore the original image from one of its corrupted versions. Recently, this field has been immensely influenced by the emergence of deep-learning techniques. Deep Image Prior (DIP) offers a new approach that forces the recovered image to be synthesized from a given deep architecture. While DIP is quite an effective unsupervised approach, it is deprecated in real-world applications because of the requirement of human assistance. In this work, we aim to find the best-recovered image without the assistance of humans by adding a stopping criterion, which will reach maximum when the iteration no longer improves the image quality. More specifically, we propose to add a pseudo noise to the corrupted image and measure the pseudo-noise component in the recovered image by the orthogonality between signal and noise. The accuracy of the orthogonal stopping criterion has been demonstrated for several tested problems such as denoising, super-resolution, and inpainting, in which 38 out of 40 experiments are higher than 95%.
With the unprecedented developments in deep learning, many methods are proposed and have achieved great success for medical image segmentation. However, unlike segmentation of natural images, most medical images such as MRI and CT are volumetric data. In order to make full use of volumetric information, 3D CNNs are widely used. However, 3D CNNs suffer from higher inference time and computation cost, which hinders their further clinical applications. Additionally, with the increased number of parameters, the risk of overfitting is higher, especially for medical images where data and annotations are expensive to acquire. To issue this problem, many 2.5D segmentation methods have been proposed to make use of volumetric spatial information with less computation cost. Despite these works lead to improvements on a variety of segmentation tasks, to the best of our knowledge, there has not previously been a large-scale empirical comparison of these methods. In this paper, we aim to present a review of the latest developments of 2.5D methods for volumetric medical image segmentation. Additionally, to compare the performance and effectiveness of these methods, we provide an empirical study of these methods on three representative segmentation tasks involving different modalities and targets. Our experimental results highlight that 3D CNNs may not always be the best choice. Besides, although all these 2.5D methods can bring performance gains to 2D baseline, not all the methods hold the benefits on different datasets. We hope the results and conclusions of our study will prove useful for the community on exploring and developing efficient volumetric medical image segmentation methods.
Deep learning models have introduced various intelligent applications to edge devices, such as image classification, speech recognition, and augmented reality. There is an increasing need of training such models on the devices in order to deliver personalized, responsive, and private learning. To address this need, this paper presents a new solution for deploying and training state-of-the-art models on the resource-constrained devices. First, the paper proposes a novel filter-pruning-based model compression method to create lightweight trainable models from large models trained in the cloud, without much loss of accuracy. Second, it proposes a novel knowledge transfer method to enable the on-device model to update incrementally in real time or near real time using incremental learning on new data and enable the on-device model to learn the unseen categories with the help of the in-cloud model in an unsupervised fashion. The results show that 1) our model compression method can remove up to 99.36% parameters of WRN-28-10, while preserving a Top-1 accuracy of over 90% on CIFAR-10; 2) our knowledge transfer method enables the compressed models to achieve more than 90% accuracy on CIFAR-10 and retain good accuracy on old categories; 3) it allows the compressed models to converge within real time (three to six minutes) on the edge for incremental learning tasks; 4) it enables the model to classify unseen categories of data (78.92% Top-1 accuracy) that it is never trained with.
We propose Wav2CLIP, a robust audio representation learning method by distilling from Contrastive Language-Image Pre-training (CLIP). We systematically evaluate Wav2CLIP on a variety of audio tasks including classification, retrieval, and generation, and show that Wav2CLIP can outperform several publicly available pre-trained audio representation algorithms. Wav2CLIP projects audio into a shared embedding space with images and text, which enables multimodal applications such as zero-shot classification, and cross-modal retrieval. Furthermore, Wav2CLIP needs just ~10% of the data to achieve competitive performance on downstream tasks compared with fully supervised models, and is more efficient to pre-train than competing methods as it does not require learning a visual model in concert with an auditory model. Finally, we demonstrate image generation from Wav2CLIP as qualitative assessment of the shared embedding space. Our code and model weights are open sourced and made available for further applications.
This study proposes a novel method to assess damages in the built environment using a deep learning workflow to quantify it. Thanks to an automated crawler, aerial images from before and after a natural disaster of 50 epicenters worldwide were obtained from Google Earth, generating a 10,000 aerial image database with a spatial resolution of 2 m per pixel. The study utilizes the algorithm Seg-Net to perform semantic segmentation of the built environment from the satellite images in both instances (prior and post-natural disasters). For image segmentation, Seg-Net is one of the most popular and general CNN architectures. The Seg-Net algorithm used reached an accuracy of 92% in the segmentation. After the segmentation, we compared the disparity between both cases represented as a percentage of change. Such coefficient of change represents the damage numerically an urban environment had to quantify the overall damage in the built environment. Such an index can give the government an estimate of the number of affected households and perhaps the extent of housing damage.
Recent advances in deep learning have facilitated near-expert medical image analysis. Supervised learning is the mainstay of current approaches, though its success requires the use of large, fully labeled datasets. However, in real-world medical practice, previously unseen disease phenotypes are encountered that have not been defined a priori in finite-size datasets. Unsupervised learning, a hypothesis-free learning framework, may play a complementary role to supervised learning. Here, we demonstrate a novel framework for voxel-wise abnormality detection in brain magnetic resonance imaging (MRI), which exploits an image reconstruction network based on an introspective variational autoencoder trained with a structural similarity constraint. The proposed network learns a latent representation for "normal" anatomical variation using a series of images that do not include annotated abnormalities. After training, the network can map unseen query images to positions in the latent space, and latent variables sampled from those positions can be mapped back to the image space to yield normal-looking replicas of the input images. Finally, the network considers abnormality scores, which are designed to reflect differences at several image feature levels, in order to locate image regions that may contain abnormalities. The proposed method is evaluated on a comprehensively annotated dataset spanning clinically significant structural abnormalities of the brain parenchyma in a population having undergone radiotherapy for brain metastasis, demonstrating that it is particularly effective for contrast-enhanced lesions, i.e., metastatic brain tumors and extracranial metastatic tumors.
Convolutional neural network (CNN)-based image denoising methods typically estimate the noise component contained in a noisy input image and restore a clean image by subtracting the estimated noise from the input. However, previous denoising methods tend to remove high-frequency information (e.g., textures) from the input. It caused by intermediate feature maps of CNN contains texture information. A straightforward approach to this problem is stacking numerous layers, which leads to a high computational cost. To achieve high performance and computational efficiency, we propose a gated texture CNN (GTCNN), which is designed to carefully exclude the texture information from each intermediate feature map of the CNN by incorporating gating mechanisms. Our GTCNN achieves state-of-the-art performance with 4.8 times fewer parameters than previous state-of-the-art methods. Furthermore, the GTCNN allows us to interactively control the texture strength in the output image without any additional modules, training, or computational costs.
Tactile sensing is inherently contact based. To use tactile data, robots need to make contact with the surface of an object. This is inefficient in applications where an agent needs to make a decision between multiple alternatives that depend the physical properties of the contact location. We propose a method to get tactile data in a non-invasive manner. The proposed method estimates the output of a tactile sensor from the depth data of the surface of the object based on past experiences. An experience dataset is built by allowing the robot to interact with various objects, collecting tactile data and the corresponding object surface depth data. We use the experience dataset to train a neural network to estimate the tactile output from depth data alone. We use GelSight tactile sensors, an image-based sensor, to generate images that capture detailed surface features at the contact location. We train a network with a dataset containing 578 tactile-image to depthmap correspondences. Given a depth-map of the surface of an object, the network outputs an estimate of the response of the tactile sensor, should it make a contact with the object. We evaluate the method with structural similarity index matrix (SSIM), a similarity metric between two images commonly used in image processing community. We present experimental results that show the proposed method outperforms a baseline that uses random images with statistical significance getting an SSIM score of 0.84 +/- 0.0056 and 0.80 +/- 0.0036, respectively.
Today deep learning is widely used for building software. A software engineering problem with deep learning is that finding an appropriate convolutional neural network (CNN) model for the task can be a challenge for developers. Recent work on AutoML, more precisely neural architecture search (NAS), embodied by tools like Auto-Keras aims to solve this problem by essentially viewing it as a search problem where the starting point is a default CNN model, and mutation of this CNN model allows exploration of the space of CNN models to find a CNN model that will work best for the problem. These works have had significant success in producing high-accuracy CNN models. There are two problems, however. First, NAS can be very costly, often taking several hours to complete. Second, CNN models produced by NAS can be very complex that makes it harder to understand them and costlier to train them. We propose a novel approach for NAS, where instead of starting from a default CNN model, the initial model is selected from a repository of models extracted from GitHub. The intuition being that developers solving a similar problem may have developed a better starting point compared to the default model. We also analyze common layer patterns of CNN models in the wild to understand changes that the developers make to improve their models. Our approach uses commonly occurring changes as mutation operators in NAS. We have extended Auto-Keras to implement our approach. Our evaluation using 8 top voted problems from Kaggle for tasks including image classification and image regression shows that given the same search time, without loss of accuracy, Manas produces models with 42.9% to 99.6% fewer number of parameters than Auto-Keras' models. Benchmarked on GPU, Manas' models train 30.3% to 641.6% faster than Auto-Keras' models.
In recent years, Generative Adversarial Networks (GANs) have improved steadily towards generating increasingly impressive real-world images. It is useful to steer the image generation process for purposes such as content creation. This can be done by conditioning the model on additional information. However, when conditioning on additional information, there still exists a large set of images that agree with a particular conditioning. This makes it unlikely that the generated image is exactly as envisioned by a user, which is problematic for practical content creation scenarios such as generating facial composites or stock photos. To solve this problem, we propose a single pipeline for text-to-image generation and manipulation. In the first part of our pipeline we introduce textStyleGAN, a model that is conditioned on text. In the second part of our pipeline we make use of the pre-trained weights of textStyleGAN to perform semantic facial image manipulation. The approach works by finding semantic directions in latent space. We show that this method can be used to manipulate facial images for a wide range of attributes. Finally, we introduce the CelebTD-HQ dataset, an extension to CelebA-HQ, consisting of faces and corresponding textual descriptions.