We present Sauron, a filter pruning method that eliminates redundant feature maps by discarding the corresponding filters with automatically-adjusted layer-specific thresholds. Furthermore, Sauron minimizes a regularization term that, as we show with various metrics, promotes the formation of feature maps clusters. In contrast to most filter pruning methods, Sauron is single-phase, similarly to typical neural network optimization, requiring fewer hyperparameters and design decisions. Additionally, unlike other cluster-based approaches, our method does not require pre-selecting the number of clusters, which is non-trivial to determine and varies across layers. We evaluated Sauron and three state-of-the-art filter pruning methods on three medical image segmentation tasks. This is an area where filter pruning has received little attention and where it can help building efficient models for medical grade computers that cannot use cloud services due to privacy considerations. Sauron achieved models with higher performance and pruning rate than the competing pruning methods. Additionally, since Sauron removes filters during training, its optimization accelerated over time. Finally, we show that the feature maps of a Sauron-pruned model were highly interpretable. The Sauron code is publicly available at https://github.com/jmlipman/SauronUNet.
Surgical instrument tracking is an active research area that can provide surgeons feedback about the location of their tools relative to anatomy. Recent tracking methods are mainly divided into two parts: segmentation and object detection. However, both can only predict 2D information, which is limiting for application to real-world surgery. An accurate 3D surgical instrument model is a prerequisite for precise predictions of the pose and depth of the instrument. Recent single-view 3D reconstruction methods are only used in natural object reconstruction and do not achieve satisfying reconstruction accuracy without 3D attribute-level supervision. Further, those methods are not suitable for the surgical instruments because of their elongated shapes. In this paper, we firstly propose an end-to-end surgical instrument reconstruction system -- Self-supervised Surgical Instrument Reconstruction (SSIR). With SSIR, we propose a multi-cycle-consistency strategy to help capture the texture information from a slim instrument while only requiring a binary instrument label map. Experiments demonstrate that our approach improves the reconstruction quality of surgical instruments compared to other self-supervised methods and achieves promising results.
In this work, we propose a generalization of the forward-forward (FF) algorithm that we call the predictive forward-forward (PFF) algorithm. Specifically, we design a dynamic, recurrent neural system that learns a directed generative circuit jointly and simultaneously with a representation circuit, combining elements of predictive coding, an emerging and viable neurobiological process theory of cortical function, with the forward-forward adaptation scheme. Furthermore, PFF efficiently learns to propagate learning signals and updates synapses with forward passes only, eliminating some of the key structural and computational constraints imposed by a backprop-based scheme. Besides computational advantages, the PFF process could be further useful for understanding the learning mechanisms behind biological neurons that make use of local (and global) signals despite missing feedback connections. We run several experiments on image data and demonstrate that the PFF procedure works as well as backprop, offering a promising brain-inspired algorithm for classifying, reconstructing, and synthesizing data patterns. As a result, our approach presents further evidence of the promise afforded by backprop-alternative credit assignment algorithms within the context of brain-inspired computing.
Previous virtual try-on methods usually focus on aligning a clothing item with a person, limiting their ability to exploit the complex pose, shape and skin color of the person, as well as the overall structure of the clothing, which is vital to photo-realistic virtual try-on. To address this potential weakness, we propose a fill in fabrics (FIFA) model, a self-supervised conditional generative adversarial network based framework comprised of a Fabricator and a unified virtual try-on pipeline with a Segmenter, Warper and Fuser. The Fabricator aims to reconstruct the clothing image when provided with a masked clothing as input, and learns the overall structure of the clothing by filling in fabrics. A virtual try-on pipeline is then trained by transferring the learned representations from the Fabricator to Warper in an effort to warp and refine the target clothing. We also propose to use a multi-scale structural constraint to enforce global context at multiple scales while warping the target clothing to better fit the pose and shape of the person. Extensive experiments demonstrate that our FIFA model achieves state-of-the-art results on the standard VITON dataset for virtual try-on of clothing items, and is shown to be effective at handling complex poses and retaining the texture and embroidery of the clothing.
Egocentric 3D human pose estimation with a single head-mounted fisheye camera has recently attracted attention due to its numerous applications in virtual and augmented reality. Existing methods still struggle in challenging poses where the human body is highly occluded or is closely interacting with the scene. To address this issue, we propose a scene-aware egocentric pose estimation method that guides the prediction of the egocentric pose with scene constraints. To this end, we propose an egocentric depth estimation network to predict the scene depth map from a wide-view egocentric fisheye camera while mitigating the occlusion of the human body with a depth-inpainting network. Next, we propose a scene-aware pose estimation network that projects the 2D image features and estimated depth map of the scene into a voxel space and regresses the 3D pose with a V2V network. The voxel-based feature representation provides the direct geometric connection between 2D image features and scene geometry, and further facilitates the V2V network to constrain the predicted pose based on the estimated scene geometry. To enable the training of the aforementioned networks, we also generated a synthetic dataset, called EgoGTA, and an in-the-wild dataset based on EgoPW, called EgoPW-Scene. The experimental results of our new evaluation sequences show that the predicted 3D egocentric poses are accurate and physically plausible in terms of human-scene interaction, demonstrating that our method outperforms the state-of-the-art methods both quantitatively and qualitatively.
In this work, we propose a novel image reconstruction framework that directly learns a neural implicit representation in k-space for ECG-triggered non-Cartesian Cardiac Magnetic Resonance Imaging (CMR). While existing methods bin acquired data from neighboring time points to reconstruct one phase of the cardiac motion, our framework allows for a continuous, binning-free, and subject-specific k-space representation.We assign a unique coordinate that consists of time, coil index, and frequency domain location to each sampled k-space point. We then learn the subject-specific mapping from these unique coordinates to k-space intensities using a multi-layer perceptron with frequency domain regularization. During inference, we obtain a complete k-space for Cartesian coordinates and an arbitrary temporal resolution. A simple inverse Fourier transform recovers the image, eliminating the need for density compensation and costly non-uniform Fourier transforms for non-Cartesian data. This novel imaging framework was tested on 42 radially sampled datasets from 6 subjects. The proposed method outperforms other techniques qualitatively and quantitatively using data from four and one heartbeat(s) and 30 cardiac phases. Our results for one heartbeat reconstruction of 50 cardiac phases show improved artifact removal and spatio-temporal resolution, leveraging the potential for real-time CMR.
Automatic defect detection for 3D printing processes, which shares many characteristics with change detection problems, is a vital step for quality control of 3D printed products. However, there are some critical challenges in the current state of practice. First, existing methods for computer vision-based process monitoring typically work well only under specific camera viewpoints and lighting situations, requiring expensive pre-processing, alignment, and camera setups. Second, many defect detection techniques are specific to pre-defined defect patterns and/or print schematics. In this work, we approach the automatic defect detection problem differently using a novel Semi-Siamese deep learning model that directly compares a reference schematic of the desired print and a camera image of the achieved print. The model then solves an image segmentation problem, identifying the locations of defects with respect to the reference frame. Unlike most change detection problems, our model is specially developed to handle images coming from different domains and is robust against perturbations in the imaging setup such as camera angle and illumination. Defect localization predictions were made in 2.75 seconds per layer using a standard MacBookPro, which is comparable to the typical tens of seconds or less for printing a single layer on an inkjet-based 3D printer, while achieving an F1-score of more than 0.9.
The VVC codec is applied to the task of multispectral image (MSI) compression using adaptive and scalable coding structures. In a 'plain' VVC approach, concepts from picture-to-picture temporal prediction are employed for decorrelation along the MSI's spectral dimension. The popular principle component analysis (PCA) for spectral decorrelation is further evaluated in combination with VVC intra-coding for spatial decorrelation. This approach is referred to as PCA-VVC. A novel adaptive MSI compression algorithm, named HPCLS, is introduced, that uses PCA and inter-prediction for spectral and VVC intra-coding for spatial decorrelation. Further, a novel adaptive scalable approach is proposed, that provides a separately decodable spectrally scaled preview of the MSI in the compressed file. Information contained in the preview is exploited in order to reduce the overall file size. All schemes are evaluated on images from the ARAD HS data set containing outdoor scenes with a high variety in brightness and color. We found that 'Plain' VVC is outperformed by both PCA-VVC and HPCLS. HPCLS shows advantageous rate-distortion (RD) behavior compared to PCA-VVC for reconstruction quality above 51dB PSNR. The performance of the scalable approach is compared to the combination of an independent RGB preview and one of HPCLS or PCA-VVC. The scalable approach shows significant benefit especially at higher preview qualities.
An increasing number of public datasets have shown a marked clinical impact on assessing anatomical structures. However, each of the datasets is small, partially labeled, and rarely investigates severe tumor subjects. Moreover, current models are limited to segmenting specific organs/tumors, which can not be extended to novel domains and classes. To tackle these limitations, we introduce embedding learned from Contrastive Language-Image Pre-training (CLIP) to segmentation models, dubbed the CLIP-Driven Universal Model. The Universal Model can better segment 25 organs and 6 types of tumors by exploiting the semantic relationship between abdominal structures. The model is developed from an assembly of 14 datasets with 3,410 CT scans and evaluated on 6,162 external CT scans from 3 datasets. We achieve the state-of-the-art results on Beyond The Cranial Vault (BTCV). Compared with dataset-specific models, the Universal Model is computationally more efficient (6x faster), generalizes better to CT scans from varying sites, and shows stronger transfer learning performance on novel tasks. The design of CLIP embedding enables the Universal Model to be easily extended to new classes without catastrophically forgetting the previously learned classes.
The recent spike in certified Artificial Intelligence (AI) tools for healthcare has renewed the debate around adoption of this technology. One thread of such debate concerns Explainable AI (XAI) and its promise to render AI devices more transparent and trustworthy. A few voices active in the medical AI space have expressed concerns on the reliability of Explainable AI techniques and especially feature attribution methods, questioning their use and inclusion in guidelines and standards. Despite valid concerns, we argue that existing criticism on the viability of post-hoc local explainability methods throws away the baby with the bathwater by generalizing a problem that is specific to image data. We begin by characterizing the problem as a lack of semantic match between explanations and human understanding. To understand when feature importance can be used reliably, we introduce a distinction between feature importance of low- and high-level features. We argue that for data types where low-level features come endowed with a clear semantics, such as tabular data like Electronic Health Records (EHRs), semantic match can be obtained, and thus feature attribution methods can still be employed in a meaningful and useful way.