Asset Search
Displaying assets 1 - 30 of 193 in total.
-
Description Qualitative saliency results comparison between different approaches on several sample images with ground truth (GT).
Article Title: Saliency Detection - FCN Salient Object Detection Using Region Cropping
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Number of changes for generating adversarial examples with defensive distillation and Temperature T. Numbers in brackets are without distillation as comparison.
Article Title: Attacks on Images - Evaluating Defensive Distillation for Defending Text Processing Neural Networks Against Adversarial Examples
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Performance comparison of different network architectures. Error bars indicate the standard error based on five repetitions of the training and testing procedure. Matrices depict results of...
Article Title: Occluded Object Recognition - Recurrent Connections Aid Occluded Object Recognition by Discounting Occluders
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The effect of WINSIZE on total perturbations in CIFAR-10.
Article Title: Attacks on Images - Improved Forward-Backward Propagation to Generate Adversarial Examples
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Proposal of variational deep embedding with regularized student-t mixture model: robustness of outliers can be given by student-t distribution; continuity in the latent space would be obtained by...
Article Title: Motion Analysis - Variational Deep Embedding with Regularized Student-t Mixture Model
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The comparsion of relative growth rate on three datasets. Comparsion one, two, and three respectively corresponds to one, two, and three columns of each dataset in this histogram. Comprison one...
Article Title: Object Detection - Referring Expression Comprehension via Co-attention and Visual Context
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Analysis of the internal representation of occluded stimuli shows discounting of occluders due to recurrent processing. (A) We define a relative distance measure to quantify if the activation of a...
Article Title: Occluded Object Recognition - Recurrent Connections Aid Occluded Object Recognition by Discounting Occluders
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Algorithm: disentangling metric.
Article Title: Generating Images - Disentangling Latent Factors of Variational Auto-encoder with Whitening
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Filing motion of an expert: (a) tool position, (b) posture of the tool.
Article Title: Motion Analysis - Comparative Research on SOM with Torus and Sphere Topologies for Peculiarity Classification of Flat Finishing Skill Training
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Visual representation of quaternions on spheres after interpolation: (a) Visual representation of interpolation using linear interpolation (red), the slerp method (green) and the squad (blue), (b)...
Article Title: Motion Analysis - Neural Network 3D Body Pose Tracking and Prediction for Motion-to-Photon Latency Compensation in Distributed Virtual Reality
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Per-category results on Cityscapes test set. Note that all the models are trained with only fine-data. Our method outperforms existing approaches on 12 out of 19 categories.
Article Title: Image Segmentation - Flow2Seg: Motion-Aided Semantic Segmentation
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Identification (a), average TSEP (b) and eye movement statistic features (c–e) for different evaluation methodologies. Error bars represent standard error of mean. (Color figure online)
Article Title: Image Denoising - Eye Movement-Based Analysis on Methodologies and Efficiency in the Process of Image Noise Evaluation
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Plot of the difference between models performance and baseline class by class when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb}...
Article Title: Generating Images - Training Discriminative Models to Evaluate Generative Ones
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Accuracy comparison for our model and sparse sampling and dense sampling.
Article Title: Object Detection - Action Recognition Based on Divide-and-Conquer
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Visualization of MA segmentation probability maps. (a) A fundus image from test set. (b) Expert annotated ground-truth. (c)–(f) Segmentation results of different methods.
Article Title: Image Segmentation - Random Drop Loss for Tiny Object Segmentation: Application to Lesion Segmentation in Fundus Images
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Overview of pixel-based dropout and filter-based dropout.
Article Title: Object Detection - Improving Reliability of Object Detection for Lunar Craters Using Monte Carlo Dropout
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The used stimuli and network models. (A) The centered target object is occluded by 2–4 digits arranged in a 3D-fashion. (B) A sketch of the four network architectures named after their connection...
Article Title: Occluded Object Recognition - Recurrent Connections Aid Occluded Object Recognition by Discounting Occluders
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The entire process of the Improved Dense Trajectories algorithm mainly consists of 3 parts: dense sampling in each spatial scale, tracking in each spatial scale separately, and trajectory...
Article Title: Object Detection - Action Recognition Based on Divide-and-Conquer
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Feature vector representation space for both models via t-SNE. One dot represents the encoding of one image in one of the five paths between the encoder and decoder. The feature representation of...
Article Title: Object Detection - Comparison Between U-Net and U-ReNet Models in OCR Tasks
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Illustration of the parallel mode and the dense mode. Note that dense mode changes the arrangement of RF units except other structures of RFB. The detail of dense mode is illustrated in (c).
Article Title: Object Detection - Dense Receptive Field Network: A Backbone Network for Object Detection
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Generated adversarial examples on the AG dataset.
Article Title: Attacks on Images - Evaluating Defensive Distillation for Defending Text Processing Neural Networks Against Adversarial Examples
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description AU names and relationships with expressions.
Article Title: Perception - Action Unit Assisted Facial Expression Recognition
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Properties of the two datasets used in this study. The real-world hand gesture dataset is relatively small, which is why we decided to complement it by a large-scale sequence classification dataset...
Article Title: Gesture Recognition - Robustness of Deep LSTM Networks in Freehand Gesture Recognition
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Transferability of 5 different methods on MNIST dataset.
Article Title: Attacks on Images - Improved Forward-Backward Propagation to Generate Adversarial Examples
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The structure of graininess-aware channel attentional module (GCAM) and graininess-aware spatial attentional module (GSAM). Note that \documentclass[12pt]{minimal} \usepackage{amsmath}...
Article Title: Image Segmentation - Attentional Residual Dense Factorized Network for Real-Time Semantic Segmentation
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Classification result of dVTS by (a) torus SOM, (b) spherical SOM: front view (left) and (c) back view(right). Colors of surrounding solid lines on plots correspond cluster number showed at left...
Article Title: Motion Analysis - Comparative Research on SOM with Torus and Sphere Topologies for Peculiarity Classification of Flat Finishing Skill Training
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Schema of investigated and our proposed models. (a) CycleGAN. (b) CycleGAN+N. (c) CycleGan+2E. (d) DesignGAN.
Article Title: Generating Images - Generative Creativity: Adversarial Learning for Bionic Design
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The PR curves of the proposed GeminiNet and 11 state-of-the-art methods on six datasets.
Article Title: Saliency Detection - Delving into the Impact of Saliency Detector: A GeminiNet for Accurate Saliency Detection
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description The evaluation results on BSDS500 dataset and NYUD+RGB dataset.
Article Title: Occluded Object Recognition - Learning Deep Structured Multi-scale Features for Crisp and Object Occlusion Edge Detection
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing -
Description Classification accuracy depending on the output frame used to generate the gesture classification.
Article Title: Gesture Recognition - Robustness of Deep LSTM Networks in Freehand Gesture Recognition
Publication Title: Artificial Neural Networks and Machine Learning – ICANN 2019: Image Processing
Displaying assets 1 - 30 of 193 in total.