Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

"Topic": models, code, and papers

Is Faster R-CNN Doing Well for Pedestrian Detection?

Jul 27, 2016
Liliang Zhang, Liang Lin, Xiaodan Liang, Kaiming He

Detecting pedestrian has been arguably addressed as a special topic beyond general object detection. Although recent deep learning object detectors such as Fast/Faster R-CNN [1, 2] have shown excellent performance for general object detection, they have limited success for detecting pedestrian, and previous leading pedestrian detectors were in general hybrid methods combining hand-crafted and deep convolutional features. In this paper, we investigate issues involving Faster R-CNN [2] for pedestrian detection. We discover that the Region Proposal Network (RPN) in Faster R-CNN indeed performs well as a stand-alone pedestrian detector, but surprisingly, the downstream classifier degrades the results. We argue that two reasons account for the unsatisfactory accuracy: (i) insufficient resolution of feature maps for handling small instances, and (ii) lack of any bootstrapping strategy for mining hard negative examples. Driven by these observations, we propose a very simple but effective baseline for pedestrian detection, using an RPN followed by boosted forests on shared, high-resolution convolutional feature maps. We comprehensively evaluate this method on several benchmarks (Caltech, INRIA, ETH, and KITTI), presenting competitive accuracy and good speed. Code will be made publicly available.

* To appear in ECCV 2016, 15 pages, 5 figures (v2: fixed some typos) 

  Access Paper or Ask Questions

Knowledge Representation in Learning Classifier Systems: A Review

Jun 12, 2015
Farzaneh Shoeleh, Mahshid Majd, Ali Hamzeh, Sattar Hashemi

Knowledge representation is a key component to the success of all rule based systems including learning classifier systems (LCSs). This component brings insight into how to partition the problem space what in turn seeks prominent role in generalization capacity of the system as a whole. Recently, knowledge representation component has received great deal of attention within data mining communities due to its impacts on rule based systems in terms of efficiency and efficacy. The current work is an attempt to find a comprehensive and yet elaborate view into the existing knowledge representation techniques in LCS domain in general and XCS in specific. To achieve the objectives, knowledge representation techniques are grouped into different categories based on the classification approach in which they are incorporated. In each category, the underlying rule representation schema and the format of classifier condition to support the corresponding representation are presented. Furthermore, a precise explanation on the way that each technique partitions the problem space along with the extensive experimental results is provided. To have an elaborated view on the functionality of each technique, a comparative analysis of existing techniques on some conventional problems is provided. We expect this survey to be of interest to the LCS researchers and practitioners since it provides a guideline for choosing a proper knowledge representation technique for a given problem and also opens up new streams of research on this topic.

  Access Paper or Ask Questions

Bayesian ensemble learning for image denoising

Aug 06, 2013
Hyuntaek Oh

Natural images are often affected by random noise and image denoising has long been a central topic in Computer Vision. Many algorithms have been introduced to remove the noise from the natural images, such as Gaussian, Wiener filtering and wavelet thresholding. However, many of these algorithms remove the fine edges and make them blur. Recently, many promising denoising algorithms have been introduced such as Non-local Means, Fields of Experts, and BM3D. In this paper, we explore Bayesian method of ensemble learning for image denoising. Ensemble methods seek to combine multiple different algorithms to retain the strengths of all methods and the weaknesses of none. Bayesian ensemble models are Non-local Means and Fields of Experts, the very successful recent algorithms. The Non-local Means presumes that the image contains an extensive amount of self-similarity. The approach of the Fields of Experts model extends traditional Markov Random Field model by learning potential functions over extended pixel neighborhoods. The two models are implemented and image denoising is performed on natural images. The experimental results obtained are used to compare with the single algorithm and discuss the ensemble learning and their approaches. Comparing to the results of Non-local Means and Fields of Experts, Ensemble learning showed improvement nearly 1dB.

* computer vision and image understanding 

  Access Paper or Ask Questions

Proceedings of the Second International Workshop on Domain-Specific Languages and Models for Robotic Systems (DSLRob 2011)

Dec 10, 2012
Ulrik Pagh Schultz, Serge Stinckwich

Proceedings of the Second International Workshop on Domain-Specific Languages and Models for Robotic Systems (DSLRob'11), held in conjunction with the 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2011), September 2011 in San Francisco, USA. The main topics of the workshop were Domain-Specific Languages (DSLs) and Model-driven Software Development (MDSD) for robotics. A domain-specific language (DSL) is a programming language dedicated to a particular problem domain that offers specific notations and abstractions that increase programmer productivity within that domain. Models offer a high-level way for domain users to specify the functionality of their system at the right level of abstraction. DSLs and models have historically been used for programming complex systems. However recently they have garnered interest as a separate field of study. Robotic systems blend hardware and software in a holistic way that intrinsically raises many crosscutting concerns (concurrency, uncertainty, time constraints, ...), for which reason, traditional general-purpose languages often lead to a poor fit between the language features and the implementation requirements. DSLs and models offer a powerful, systematic way to overcome this problem, enabling the programmer to quickly and precisely implement novel software solutions to complex problems

* Index submission 

  Access Paper or Ask Questions

Exploratory simulation of an Intelligent Iris Verifier Distributed System

Jun 18, 2011
Nicolaie Popescu-Bodorin, Valentina E. Balas

This paper discusses some topics related to the latest trends in the field of evolutionary approaches to iris recognition. It presents the results of an exploratory experimental simulation whose goal was to analyze the possibility of establishing an Interchange Protocol for Digital Identities evolved in different geographic locations interconnected through and into an Intelligent Iris Verifier Distributed System (IIVDS) based on multi-enrollment. Finding a logically consistent model for the Interchange Protocol is the key factor in designing the future large-scale iris biometric networks. Therefore, the logical model of such a protocol is also investigated here. All tests are made on Bath Iris Database and prove that outstanding power of discrimination between the intra- and the inter-class comparisons can be achieved by an IIVDS, even when practicing 52.759.182 inter-class and 10.991.943 intra-class comparisons. Still, the test results confirm that inconsistent enrollment can change the logic of recognition from a fuzzified 2-valent consistent logic of biometric certitudes to a fuzzified 3-valent inconsistent possibilistic logic of biometric beliefs justified through experimentally determined probabilities, or to a fuzzified 8-valent logic which is almost consistent as a biometric theory - this quality being counterbalanced by an absolutely reasonable loss in the user comfort level.

* Proc. 6th IEEE International Symposium on Applied Computational Intelligence and Informatics, pp. 259 - 262, IEEE Press, June 2011 
* 4 pages, 2 figures, latest version: 

  Access Paper or Ask Questions

TranSiam: Fusing Multimodal Visual Features Using Transformer for Medical Image Segmentation

Apr 26, 2022
Xuejian Li, Shiqiang Ma, Jijun Tang, Fei Guo

Automatic segmentation of medical images based on multi-modality is an important topic for disease diagnosis. Although the convolutional neural network (CNN) has been proven to have excellent performance in image segmentation tasks, it is difficult to obtain global information. The lack of global information will seriously affect the accuracy of the segmentation results of the lesion area. In addition, there are visual representation differences between multimodal data of the same patient. These differences will affect the results of the automatic segmentation methods. To solve these problems, we propose a segmentation method suitable for multimodal medical images that can capture global information, named TranSiam. TranSiam is a 2D dual path network that extracts features of different modalities. In each path, we utilize convolution to extract detailed information in low level stage, and design a ICMT block to extract global information in high level stage. ICMT block embeds convolution in the transformer, which can extract global information while retaining spatial and detailed information. Furthermore, we design a novel fusion mechanism based on cross attention and selfattention, called TMM block, which can effectively fuse features between different modalities. On the BraTS 2019 and BraTS 2020 multimodal datasets, we have a significant improvement in accuracy over other popular methods.

  Access Paper or Ask Questions

DirecFormer: A Directed Attention in Transformer Approach to Robust Action Recognition

Mar 19, 2022
Thanh-Dat Truong, Quoc-Huy Bui, Chi Nhan Duong, Han-Seok Seo, Son Lam Phung, Xin Li, Khoa Luu

Human action recognition has recently become one of the popular research topics in the computer vision community. Various 3D-CNN based methods have been presented to tackle both the spatial and temporal dimensions in the task of video action recognition with competitive results. However, these methods have suffered some fundamental limitations such as lack of robustness and generalization, e.g., how does the temporal ordering of video frames affect the recognition results? This work presents a novel end-to-end Transformer-based Directed Attention (DirecFormer) framework for robust action recognition. The method takes a simple but novel perspective of Transformer-based approach to understand the right order of sequence actions. Therefore, the contributions of this work are three-fold. Firstly, we introduce the problem of ordered temporal learning issues to the action recognition problem. Secondly, a new Directed Attention mechanism is introduced to understand and provide attentions to human actions in the right order. Thirdly, we introduce the conditional dependency in action sequence modeling that includes orders and classes. The proposed approach consistently achieves the state-of-the-art (SOTA) results compared with the recent action recognition methods, on three standard large-scale benchmarks, i.e. Jester, Kinetics-400 and Something-Something-V2.

* Accepted to CVPR 2022 

  Access Paper or Ask Questions

Multi-Stage Prompting for Knowledgeable Dialogue Generation

Mar 16, 2022
Zihan Liu, Mostofa Patwary, Ryan Prenger, Shrimai Prabhumoye, Wei Ping, Mohammad Shoeybi, Bryan Catanzaro

Existing knowledge-grounded dialogue systems typically use finetuned versions of a pretrained language model (LM) and large-scale knowledge bases. These models typically fail to generalize on topics outside of the knowledge base, and require maintaining separate potentially large checkpoints each time finetuning is needed. In this paper, we aim to address these limitations by leveraging the inherent knowledge stored in the pretrained LM as well as its powerful generation ability. We propose a multi-stage prompting approach to generate knowledgeable responses from a single pretrained LM. We first prompt the LM to generate knowledge based on the dialogue context. Then, we further prompt it to generate responses based on the dialogue context and the previously generated knowledge. Results show that our knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness. In addition, our multi-stage prompting outperforms the finetuning-based dialogue model in terms of response knowledgeability and engagement by up to 10% and 5%, respectively. Furthermore, we scale our model up to 530 billion parameters and show that larger LMs improve the generation correctness score by up to 10%, and response relevance, knowledgeability and engagement by up to 10%. Our code is available at:

  Access Paper or Ask Questions

Identifiability of Causal-based Fairness Notions: A State of the Art

Mar 11, 2022
Karima Makhlouf, Sami Zhioua, Catuscia Palamidessi

Machine learning algorithms can produce biased outcome/prediction, typically, against minorities and under-represented sub-populations. Therefore, fairness is emerging as an important requirement for the large scale application of machine learning based technologies. The most commonly used fairness notions (e.g. statistical parity, equalized odds, predictive parity, etc.) are observational and rely on mere correlation between variables. These notions fail to identify bias in case of statistical anomalies such as Simpson's or Berkson's paradoxes. Causality-based fairness notions (e.g. counterfactual fairness, no-proxy discrimination, etc.) are immune to such anomalies and hence more reliable to assess fairness. The problem of causality-based fairness notions, however, is that they are defined in terms of quantities (e.g. causal, counterfactual, and path-specific effects) that are not always measurable. This is known as the identifiability problem and is the topic of a large body of work in the causal inference literature. This paper is a compilation of the major identifiability results which are of particular relevance for machine learning fairness. The results are illustrated using a large number of examples and causal graphs. The paper would be of particular interest to fairness researchers, practitioners, and policy makers who are considering the use of causality-based fairness notions as it summarizes and illustrates the major identifiability results

  Access Paper or Ask Questions

A Preliminary Study on Aging Examining Online Handwriting

Mar 08, 2022
Marcos Faundez-Zanuy, Enric Sesa-Nogueras, Josep Roure-Alcobé, Anna Esposito, Jiri Mekyska, Karmele López-de-Ipiña

In order to develop infocommunications devices so that the capabilities of the human brain may interact with the capabilities of any artificially cognitive system a deeper knowledge of aging is necessary. Especially if society does not want to exclude elder people and wants to develop automatic systems able to help and improve the quality of life of this group of population, healthy individuals as well as those with cognitive decline or other pathologies. This paper tries to establish the variations in handwriting tasks with the goal to obtain a better knowledge about aging. We present the correlation results between several parameters extracted from online handwriting and the age of the writers. It is based on BIOSECURID database, which consists of 400 people that provided several biometric traits, including online handwriting. The main idea is to identify those parameters that are more stable and those more age dependent. One challenging topic for disease diagnose is the differentiation between healthy and pathological aging. For this purpose, it is necessary to be aware of handwriting parameters that are, in general, not affected by aging and those who experiment changes, increase or decrease their values, because of it. This paper contributes to this research line analyzing a selected set of online handwriting parameters provided by a healthy group of population aged from 18 to 70 years. Preliminary results show that these parameters are not affected by aging and therefore, changes in their values can only be attributed to motor or cognitive disorders.

* 2014 5th IEEE Conference on Cognitive Infocommunications (CogInfoCom), 2014, pp. 221-224 
* 4 pages 

  Access Paper or Ask Questions