7,219 Hits in 4.3 sec

Grid Loss: Detecting Occluded Faces [article]

Michael Opitz, Georg Waltner, Georg Poier, Horst Possegger, Horst Bischof
2016 arXiv   pre-print
To address this issue, we propose a novel loss layer for CNNs, named grid loss, which minimizes the error rate on sub-blocks of a convolution layer independently rather than over the whole feature map.  ...  Detection of partially occluded objects is a challenging computer vision problem.  ...  of grid loss for detecting occluded objects.  ... 
arXiv:1609.00129v1 fatcat:atkusojtivejth6enpk6t4ckn4

Deep Object Tracking on Dynamic Occupancy Grid Maps Using RNNs [article]

Nico Engel, Stefan Hoermann, Philipp Henzler, Klaus Dietmayer
2018 arXiv   pre-print
Furthermore, we gain an improved detection for occluded objects and a more consistent size estimation due to the usage of time series as input and the memory about previous states introduced by the recurrent  ...  Multiple laser scanners are fused into a dynamic occupancy grid map resulting in a 360 perception of the environment.  ...  With only the CNN (top row), the network fails to detect the rear object when it gets occluded in time step 2, even though it detected both objects in the first frame.  ... 
arXiv:1805.08986v1 fatcat:ejettusx45e47ezptgvhmycmli

Capturing Object Detection Uncertainty in Multi-Layer Grid Maps [article]

Sascha Wirges and Marcel Reith-Braun and Martin Lauer and Christoph Stiller
2019 arXiv   pre-print
We train our models based on the KITTI object detection data set.  ...  The input consists of a multi-layer grid map which is well-suited for sensor fusion, free-space estimation and machine learning.  ...  To analyze the influence of the loss function we further use a variant which uses an L1-Loss (as presented in [6] ) instead of an L2-Loss, named Aleatoric L1.  ... 
arXiv:1901.11284v1 fatcat:hualsezxjbab5cv5mfis33jzgq

LapNet : Automatic Balanced Loss and Optimal Assignment for Real-Time Dense Object Detection [article]

Florian Chabot, Quoc-Cuong Pham, Mohamed Chaouch
2020 arXiv   pre-print
It needs to face several challenges: small object detection, heavily occluded objects and real time processing.  ...  In fact, the objective of anchor weights is to balance the loss of each anchor grid.  ... 
arXiv:1911.01149v2 fatcat:po3cuvltdbgujpydfrpa7ew3yq

Object Detection on Dynamic Occupancy Grid Maps Using Deep Learning and Automatic Label Generation [article]

Stefan Hoermann, Philipp Henzler, Martin Bach, Klaus Dietmayer
2018 arXiv   pre-print
Therefore, we propose a loss function counteracting the high imbalance between mostly static background and extremely rare dynamic grid cells.  ...  We tackle the problem of object detection and pose estimation in a shared space downtown environment.  ...  Detecting objects in grid maps widely decouples sensor fusion from object tracking, i.e. performing object detection after dynamic grid mapping.  ... 
arXiv:1802.02202v1 fatcat:zr6wpyaqojeg3c5pmeeld6uapy

Loss and reappearance of gap junctions in regenerating liver

A. G. Yee
1978 Journal of Cell Biology  
Although more difficult to detect, the small gap junctions were also observed on the corresponding E-fracture face; they are characterized as a cluster of some five to six pits or depressions in polygonal  ...  The small arrows indicate regions in which the ridges of the P-fracture face appear to be co-linear with the grooves on the E-fracture face.  ... 
doi:10.1083/jcb.78.2.554 pmid:690179 pmcid:PMC2110135 fatcat:auhcr5zkuzf3bjk4avtdkxclbi

Dynamic Occupancy Grid Prediction for Urban Autonomous Driving: A Deep Learning Approach with Fully Automatic Labeling [article]

Stefan Hoermann and Martin Bach and Klaus Dietmayer
2017 arXiv   pre-print
More specifically, a dynamic occupancy grid map is utilized as input to a deep convolutional neural network.  ...  Pixel-wise balancing is applied in the loss function counteracting the extreme imbalance between static and dynamic cells.  ...  Predicting Partially Occluded Objects Another pattern, typical for dynamic occupancy grids, is that for objects entering the field of view, the area estimated as occupied tends to grow.  ... 
arXiv:1705.08781v2 fatcat:6s7wt5mcyzcgrho3q7kzizfcsu

A Grid-based Representation for Human Action Recognition [article]

Soufiane Lamghari, Guillaume-Alexandre Bilodeau, Nicolas Saunier
2020 arXiv   pre-print
novel method for human action recognition that encodes efficiently the most discriminative appearance information of an action with explicit attention on representative pose features, into a new compact grid  ...  Our GRAR (Grid-based Representation for Action Recognition) method is tested on several benchmark datasets demonstrating that our model can accurately recognize human actions, despite intra-class appearance  ...  (b) and (c) show occluded humans handled with our explicit attention on pose. Fig. 3 . 3 Example of a grid image from the CAE dataset.  ... 
arXiv:2010.08841v2 fatcat:iublzset45gxpdsmukvsedptjm

A Grid-Based Framework for Collective Perception in Autonomous Vehicles

Jorge Godoy, Víctor Jiménez, Antonio Artuñedo, Jorge Villagra
2021 Sensors  
To that end, the environment is modeled using an occupancy grid where occupied, and free and uncertain space is considered.  ...  For each sensor, including V2X, independent grids are calculated from sensor measurements and uncertainties and then fused in terms of both occupancy and confidence.  ...  ., radars have a better performance than cameras when facing rain or fog, and cameras are much more sensitive to lighting conditions.  ... 
doi:10.3390/s21030744 pmid:33499331 fatcat:2k246i46rbfcjdr6s3fzicvr2i

FH-YOLOv4 with Constrained Aspect Ratio Loss for Video Face Detection and Public Safety

Yue Wang, Liang Hong, Dewen Gu, Pingping Fu, Wei Zhang
2022 Discrete Dynamics in Nature and Society  
Video face detection is a crucial first step in many facial recognition and face analysis systems.  ...  In this paper, we first introduce the constrained aspect ratio loss (CARLoss) for better facial boxes regression and incorporate it into the modified FH-YOLOv4, then the IoU Tracker-based video face image  ...  For instance, the mature detection framework YOLOv4 [14] uses the architecture of CSPDarkNet53 with an SPP layer as backbone, PANet as Neck and YOLO detection head. e detection of occluded targets, small  ... 
doi:10.1155/2022/8355174 fatcat:bew3lm7kw5fmfejwhg6oods76a

Multivehicle Cooperative Local Mapping: A Methodology Based on Occupancy Grid Map Merging

Hao Li, Manabu Tsukada, Fawzi Nashashibi, Michel Parent
2014 IEEE transactions on intelligent transportation systems (Print)  
We present a variety of experiments that validate the effectiveness of the proposed occupancy grid map merging method.  ...  In this paper, we introduce a method of occupancy grid map merging, dedicated to multi-vehicle cooperative local mapping purpose in outdoor environment.  ...  Let M A(t) and M B(t) denote their occupancy grid maps. Without loss of generality, we neglect time index t in the following. In this section, we describe our method for merging M A and M B . A.  ... 
doi:10.1109/tits.2014.2309639 fatcat:d27lukkj7zexxlep6w2w4f74l4

Optimizing Nondecomposable Loss Functions in Structured Prediction

Mani Ranjbar, Tian Lan, Yang Wang, Steven N. Robinovitch, Ze-Nian Li, Greg Mori
2013 IEEE Transactions on Pattern Analysis and Machine Intelligence  
The loss augmented inference forms a Quadratic Program (QP), which we solve using LP relaxation.  ...  We attack this optimization problem by approximating the loss function with a piecewise linear function.  ...  Based on the ordering between the detections the occluded pixels are ignored for appearance model estimation.  ... 
doi:10.1109/tpami.2012.168 pmid:22868650 pmcid:PMC3547074 fatcat:etkqq47cjfctdd4cmnrkgddshe

Structure of Crenezumab Complex with Aβ Shows Loss of β-Hairpin

Mark Ultsch, Bing Li, Till Maurer, Mary Mathieu, Oskar Adolfsson, Andreas Muhs, Andrea Pfeifer, Maria Pihlgren, Travis W. Bainbridge, Mike Reichelt, James A. Ernst, Charles Eigenbrot (+4 others)
2016 Scientific Reports  
The aromatic side chains of Phe19 Aβ and Phe20 Aβ formed a π -π stacking network comprised of one "face-to-face" and two "face-to-edge" interactions with light chain residues Trp96 LC and His34 LC (Fig  ...  We first incubated crenezumab with Aβ 1-42 fibrils on TEM grids and then labeled the bound crenezumab with a secondary biotinylated antibody, which was subsequently detected with streptavidin conjugated  ... 
doi:10.1038/srep39374 pmid:27996029 pmcid:PMC5171940 fatcat:ye5btdwitvg5zbzdysvslot2ai

Dual Grid Net: hand mesh vertex regression from single depth maps [article]

Chengde Wan, Thomas Probst, Luc Van Gool, Angela Yao
2019 arXiv   pre-print
In the first stage, the network estimates a dense correspondence field for every pixel on the depth map or image grid to the mesh grid.  ...  Finally, we sample from the mesh grid to recover the mesh vertices, and fit it an articulated template mesh in closed form.  ...  (a) (b) (c) (d) Mapping from image grid to mesh grid In this section, we show how to recover all mesh vertices, including occluded ones, from the estimated per-pixel mesh coordinate and features on the  ... 
arXiv:1907.10695v1 fatcat:67ln42zlujf73kgiwddedplruu

Incorporating shading losses in solar photovoltaic potential assessment at the municipal scale

Ha T. Nguyen, Joshua M. Pearce
2012 Solar Energy  
The output is a boolean shading raster where a pixel receives a score of 1 if it is occluded by any other pixel and 0 otherwise.  ...  The remaining error-free data was fit into different planar faces, the procedure of which is explained in Figure 2b .  ... 
doi:10.1016/j.solener.2012.01.017 fatcat:hoenwrsh2ba4zla65qr5jmwylq
« Previous Showing results 1 — 15 out of 7,219 results