Unusual scene detection using distributed behaviour model and sparse representation

The ability to detect unusual events in surviellance footage as they happen is a highly desireable feature for a surveillance system. However, this problem remains challenging in crowded scenes due to occlusions and the clustering of people. In this paper, we propose using the Distributed Behavior Model (DBM), which has been widely used in computer graphics, for video event detection. Our approach does not rely on object tracking, and is robust to camera movements. We use sparse coding for classification, and test our approach on various datasets. Our proposed approach outperforms a state-of-the-art work which uses the social force model and Latent Dirichlet Allocation.

Lighting invariant urban street classification

In this paper we propose the hybrid use of illuminant invariant and RGB images to perform image classification of urban scenes despite challenging variation in lighting conditions. Coping with lighting change (and the shadows thereby invoked) is a non-negotiable requirement for long term autonomy using vision. One aspect of this is the ability to reliably classify scene components in the presence of marked and often sudden changes in lighting. This is the focus of this paper. Posed with the task of classifying all parts in a scene from a full colour image, we propose that lighting invariant transforms can reduce the variability of the scene, resulting in a more reliable classification. We leverage the ideas of “data transfer” for classification, beginning with full colour images for obtaining candidate scene-level matches using global image descriptors. This is commonly followed by superpixellevel matching with local features. However, we show that if the RGB images are subjected to an illuminant invariant transform before computing the superpixel-level features, classification is significantly more robust to scene illumination effects. The approach is evaluated using three datasets. The first being our own dataset and the second being the KITTI dataset using manually generated ground truth for quantitative analysis. We qualitatively evaluate the method on a third custom dataset over a 750m trajectory.

An application of fuzzy DL-Based semantic perception to soil container classification

Semantic perception and object labeling are key requirements for robots interacting with objects on a higher level. Symbolic annotation of objects allows the usage of planning algorithms for object interaction, for instance in a typical fetchand-carry scenario. In current research, perception is usually based on 3D scene reconstruction and geometric model matching, where trained features are matched with a 3D sample point cloud. In this work we propose a semantic perception method which is based on spatio-semantic features. These features are defined in a natural, symbolic way, such as geometry and spatial relation. In contrast to point-based model matching methods, a spatial ontology is used where objects are rather described how they "look like", similar to how a human would described unknown objects to another person. A fuzzy based reasoning approach matches perceivable features with a spatial ontology of the objects. The approach provides a method which is able to deal with senor noise and occlusions. Another advantage is that no training phase is needed in order to learn object features. The use-case of the proposed method is the detection of soil sample containers in an outdoor environment which have to be collected by a mobile robot. The approach is verified using real world experiments.

Using multi-label classification for acoustic pattern detection and assisting bird species surveys

Acoustics is a rich source of environmental information that can reflect the ecological dynamics. To deal with the escalating acoustic data, a variety of automated classification techniques have been used for acoustic patterns or scene recognition, including urban soundscapes such as streets and restaurants; and natural soundscapes such as raining and thundering. It is common to classify acoustic patterns under the assumption that a single type of soundscapes present in an audio clip. This assumption is reasonable for some carefully selected audios. However, only few experiments have been focused on classifying simultaneous acoustic patterns in long-duration recordings. This paper proposes a binary relevance based multi-label classification approach to recognise simultaneous acoustic patterns in one-minute audio clips. By utilising acoustic indices as global features and multilayer perceptron as a base classifier, we achieve good classification performance on in-the-field data. Compared with single-label classification, multi-label classification approach provides more detailed information about the distributions of various acoustic patterns in long-duration recordings. These results will merit further biodiversity investigations, such as bird species surveys.

Recognition of Kannada characters extracted from scene images

In this paper, we describe a method for feature extraction and classification of characters manually isolated from scene or natural images. Characters in a scene image may be affected by low resolution, uneven illumination or occlusion. We propose a novel method to perform binarization on gray scale images by minimizing energy functional. Discrete Cosine Transform and Angular Radial Transform are used to extract the features from characters after normalization for scale and translation. We have evaluated our method on the complete test set of Chars74k dataset for English and Kannada scripts consisting of handwritten and synthesized characters, as well as characters extracted from camera captured images. We utilize only synthesized and handwritten characters from this dataset as training set. Nearest neighbor classification is used in our experiments.

An adaptation framework for head-pose classification in dynamic multi-view scenarios

Multi-view head-pose estimation in low-resolution, dynamic scenes is difficult due to blurred facial appearance and perspective changes as targets move around freely in the environment. Under these conditions, acquiring sufficient training examples to learn the dynamic relationship between position, face appearance and head-pose can be very expensive. Instead, a transfer learning approach is proposed in this work. Upon learning a weighted-distance function from many examples where the target position is fixed, we adapt these weights to the scenario where target positions are varying. The adaptation framework incorporates reliability of the different face regions for pose estimation under positional variation, by transforming the target appearance to a canonical appearance corresponding to a reference scene location. Experimental results confirm effectiveness of the proposed approach, which outperforms state-of-the-art by 9.5% under relevant conditions. To aid further research on this topic, we also make DPOSE- a dynamic, multi-view head-pose dataset with ground-truth publicly available with this paper.

EVA: Endoscopic Video Analysis of the surgical scene for the assessment of MIS psychomotor skills

The present work covers the first validation efforts of the EVA Tracking System for the assessment of minimally invasive surgery (MIS) psychomotor skills. Instrument movements were recorded for 42 surgeons (4 expert, 22 residents, 16 novice medical students) and analyzed for a box trainer peg transfer task. Construct validation was established for 7/9 motion analysis parameters (MAPs). Concurrent validation was determined for 8/9 MAPs against the TrEndo Tracking System. Finally, automatic determination of surgical proficiency based on the MAPs was sought by 3 different approaches to supervised classification (LDA, SVM, ANFIS), with accuracy results of 61.9%, 83.3% and 80.9% respectively. Results not only reflect on the validation of EVA for skills? assessment, but also on the relevance of motion analysis of instruments in the determination of surgical competence.

A new unsupervised convolutional neural network model for Chinese scene text detection

As one of the most popular deep learning models, convolution neural network (CNN) has achieved huge success in image information extraction. Traditionally CNN is trained by supervised learning method with labeled data and used as a classifier by adding a classification layer in the end. Its capability of extracting image features is largely limited due to the difficulty of setting up a large training dataset. In this paper, we propose a new unsupervised learning CNN model, which uses a so-called convolutional sparse auto-encoder (CSAE) algorithm pre-Train the CNN. Instead of using labeled natural images for CNN training, the CSAE algorithm can be used to train the CNN with unlabeled artificial images, which enables easy expansion of training data and unsupervised learning. The CSAE algorithm is especially designed for extracting complex features from specific objects such as Chinese characters. After the features of articficial images are extracted by the CSAE algorithm, the learned parameters are used to initialize the first CNN convolutional layer, and then the CNN model is fine-Trained by scene image patches with a linear classifier. The new CNN model is applied to Chinese scene text detection and is evaluated with a multilingual image dataset, which labels Chinese, English and numerals texts separately. More than 10% detection precision gain is observed over two CNN models.