Deep Lesion Graph in the Wild: Relationship Learning and Organization of Significant Radiology Image Findings in a Diverse Large-Scale Lesion Database

Author(s):  
Ke Yan ◽  
Xiaosong Wang ◽  
Le Lu ◽  
Ling Zhang ◽  
Adam P. Harrison ◽  
...  
Author(s):  
Mehdi Bahri ◽  
Eimear O’ Sullivan ◽  
Shunwang Gong ◽  
Feng Liu ◽  
Xiaoming Liu ◽  
...  

AbstractStandard registration algorithms need to be independently applied to each surface to register, following careful pre-processing and hand-tuning. Recently, learning-based approaches have emerged that reduce the registration of new scans to running inference with a previously-trained model. The potential benefits are multifold: inference is typically orders of magnitude faster than solving a new instance of a difficult optimization problem, deep learning models can be made robust to noise and corruption, and the trained model may be re-used for other tasks, e.g. through transfer learning. In this paper, we cast the registration task as a surface-to-surface translation problem, and design a model to reliably capture the latent geometric information directly from raw 3D face scans. We introduce Shape-My-Face (SMF), a powerful encoder-decoder architecture based on an improved point cloud encoder, a novel visual attention mechanism, graph convolutional decoders with skip connections, and a specialized mouth model that we smoothly integrate with the mesh convolutions. Compared to the previous state-of-the-art learning algorithms for non-rigid registration of face scans, SMF only requires the raw data to be rigidly aligned (with scaling) with a pre-defined face template. Additionally, our model provides topologically-sound meshes with minimal supervision, offers faster training time, has orders of magnitude fewer trainable parameters, is more robust to noise, and can generalize to previously unseen datasets. We extensively evaluate the quality of our registrations on diverse data. We demonstrate the robustness and generalizability of our model with in-the-wild face scans across different modalities, sensor types, and resolutions. Finally, we show that, by learning to register scans, SMF produces a hybrid linear and non-linear morphable model. Manipulation of the latent space of SMF allows for shape generation, and morphing applications such as expression transfer in-the-wild. We train SMF on a dataset of human faces comprising 9 large-scale databases on commodity hardware.


Author(s):  
Matthias Müller ◽  
Adel Bibi ◽  
Silvio Giancola ◽  
Salman Alsubaihi ◽  
Bernard Ghanem

Development ◽  
1996 ◽  
Vol 123 (1) ◽  
pp. 241-254 ◽  
Author(s):  
T.T. Whitfield ◽  
M. Granato ◽  
F.J. van Eeden ◽  
U. Schach ◽  
M. Brand ◽  
...  

Mutations giving rise to anatomical defects in the inner ear have been isolated in a large scale screen for mutations causing visible abnormalities in the zebrafish embryo (Haffter, P., Granato, M., Brand, M. et al. (1996) Development 123, 1–36). 58 mutants have been classified as having a primary ear phenotype; these fall into several phenotypic classes, affecting presence or size of the otoliths, size and shape of the otic vesicle and formation of the semicircular canals, and define at least 20 complementation groups. Mutations in seven genes cause loss of one or both otoliths, but do not appear to affect development of other structures within the ear. Mutations in seven genes affect morphology and patterning of the inner ear epithelium, including formation of the semicircular canals and, in some, development of sensory patches (maculae and cristae). Within this class, dog-eared mutants show abnormal development of semicircular canals and lack cristae within the ear, while in van gogh, semicircular canals fail to form altogether, resulting in a tiny otic vesicle containing a single sensory patch. Both these mutants show defects in the expression of homeobox genes within the otic vesicle. In a further class of mutants, ear size is affected while patterning appears to be relatively normal; mutations in three genes cause expansion of the otic vesicle, while in little ears and microtic, the ear is abnormally small, but still contains all five sensory patches, as in the wild type. Many of the ear and otolith mutants show an expected behavioural phenotype: embryos fail to balance correctly, and may swim on their sides, upside down, or in circles. Several mutants with similar balance defects have also been isolated that have no obvious structural ear defect, but that may include mutants with vestibular dysfunction of the inner ear (Granato, M., van Eeden, F. J. M., Schach, U. et al. (1996) Development, 123, 399–413,). Mutations in 19 genes causing primary defects in other structures also show an ear defect. In particular, ear phenotypes are often found in conjunction with defects of neural crest derivatives (pigment cells and/or cartilaginous elements of the jaw). At least one mutant, dog-eared, shows defects in both the ear and another placodally derived sensory system, the lateral line, while hypersensitive mutants have additional trunk lateral line organs.


Diversity ◽  
2020 ◽  
Vol 12 (8) ◽  
pp. 314 ◽  
Author(s):  
Pedro Romero-Vidal ◽  
Fernando Hiraldo ◽  
Federica Rosseto ◽  
Guillermo Blanco ◽  
Martina Carrete ◽  
...  

Illegal wildlife trade, which mostly focuses on high-demand species, constitutes a major threat to biodiversity. However, whether poaching is an opportunistic crime within high-demand taxa such as parrots (i.e., harvesting proportional to species availability in the wild), or is selectively focused on particular, more desirable species, is still under debate. Answering this question has important conservation implications because selective poaching can lead to the extinction of some species through overharvesting. However, the challenges of estimating species abundances in the wild have hampered studies on this subject. We conducted a large-scale survey in Colombia to simultaneously estimate the relative abundance of wild parrots through roadside surveys (recording 10,811 individuals from 25 species across 2221 km surveyed) and as household, illegally trapped pets in 282 sampled villages (1179 individuals from 21 species). We used for the first time a selectivity index to test selection on poaching. Results demonstrated that poaching is not opportunistic, but positively selects species based on their attractiveness, defined as a function of species size, coloration, and ability to talk, which is also reflected in their local prices. Our methodological approach, which shows how selection increases the conservation impacts of poaching for parrots, can be applied to other taxa also impacted by harvesting for trade or other purposes.


Science ◽  
2020 ◽  
Vol 369 (6500) ◽  
pp. 194-197 ◽  
Author(s):  
Lee Harten ◽  
Amitay Katz ◽  
Aya Goldshtein ◽  
Michal Handel ◽  
Yossi Yovel

How animals navigate over large-scale environments remains a riddle. Specifically, it is debated whether animals have cognitive maps. The hallmark of map-based navigation is the ability to perform shortcuts, i.e., to move in direct but novel routes. When tracking an animal in the wild, it is extremely difficult to determine whether a movement is truly novel because the animal’s past movement is unknown. We overcame this difficulty by continuously tracking wild fruit bat pups from their very first flight outdoors and over the first months of their lives. Bats performed truly original shortcuts, supporting the hypothesis that they can perform large-scale map-based navigation. We documented how young pups developed their visual-based map, exemplifying the importance of exploration and demonstrating interindividual differences.


Symmetry ◽  
2020 ◽  
Vol 12 (11) ◽  
pp. 1832
Author(s):  
Tomasz Hachaj ◽  
Patryk Mazurek

Deep learning-based feature extraction methods and transfer learning have become common approaches in the field of pattern recognition. Deep convolutional neural networks trained using tripled-based loss functions allow for the generation of face embeddings, which can be directly applied to face verification and clustering. Knowledge about the ground truth of face identities might improve the effectiveness of the final classification algorithm; however, it is also possible to use ground truth clusters previously discovered using an unsupervised approach. The aim of this paper is to evaluate the potential improvement of classification results of state-of-the-art supervised classification methods trained with and without ground truth knowledge. In this study, we use two sufficiently large data sets containing more than 200,000 “taken in the wild” images, each with various resolutions, visual quality, and face poses which, in our opinion, guarantee the statistical significance of the results. We examine several clustering and supervised pattern recognition algorithms and find that knowledge about the ground truth has a very small influence on the Fowlkes–Mallows score (FMS) of the classification algorithm. In the case of the classification algorithm that obtained the highest accuracy in our experiment, the FMS improved by only 5.3% (from 0.749 to 0.791) in the first data set and by 6.6% (from 0.652 to 0.718) in the second data set. Our results show that, beside highly secure systems in which face verification is a key component, face identities discovered by unsupervised approaches can be safely used for training supervised classifiers. We also found that the Silhouette Coefficient (SC) of unsupervised clustering is positively correlated with the Adjusted Rand Index, V-measure score, and Fowlkes–Mallows score and, so, we can use the SC as an indicator of clustering performance when the ground truth of face identities is not known. All of these conclusions are important findings for large-scale face verification problems. The reason for this is the fact that skipping the verification of people’s identities before supervised training saves a lot of time and resources.


2010 ◽  
Vol 25 (9) ◽  
pp. 520-529 ◽  
Author(s):  
Linda Laikre ◽  
Michael K. Schwartz ◽  
Robin S. Waples ◽  
Nils Ryman

Sign in / Sign up

Export Citation Format

Share Document