Martin Hofmann

Background segmentation with feedback: The Pixel-Based Adaptive Segmenter

In this paper we present a novel method for foreground segmentation. Our proposed approach follows a non-parametric background modeling paradigm, thus the background is modeled by a history of recently observed pixel values. The foreground decision depends on a decision threshold. The background update is based on a learning parameter. We extend both of these parameters to dynamic per-pixel state variables and introduce dynamic controllers for each of them. Furthermore, both controllers are steered by an estimate of the background dynamics. In our experiments, the proposed Pixel-Based Adaptive Segmenter (PBAS) outperforms most state-of-the-art methods.

The TUM Gait from Audio, Image and Depth (GAID) database

HighlightsPresentation of the new freely available TUM Gait from Audio, Image and Depth (GAID) database.Advancing gait based person identification by multimodal feature extraction.Gait based recognition of person traits: gender, age, height, shoe type.Baseline results and fusion for gait recognition using RGB, depth and audio. Recognizing people by the way they walk-also known as gait recognition-has been studied extensively in the recent past. Recent gait recognition methods solely focus on data extracted from an RGB video stream. With this work, we provide a means for multimodal gait recognition, by introducing the freely available TUM Gait from Audio, Image and Depth (GAID) database. This database simultaneously contains RGB video, depth and audio. With 305 people in three variations, it is one of the largest to-date. To further investigate challenges of time variation, a subset of 32 people is recorded a second time. We define standardized experimental setups for both person identification and for the assessment of the soft biometrics age, gender, height, and shoe type. For all defined experiments, we present several baseline results on all available modalities. These effectively demonstrate multimodal fusion being beneficial to gait recognition.

2.5D gait biometrics using the Depth Gradient Histogram Energy Image

Using gait recognition methods, people can be identified by the way they walk. The most successful and efficient of these methods are based on the Gait Energy Image (GEI). In this paper, we extend the traditional Gait Energy Image by including depth information. First, GEI is extended by calculating the required silhouettes using depth data. We then formulate a completely new feature, which we call the Depth Gradient Histogram Energy Image (DGHEI). We compare the improved depth-GEI and the new DGHEI to the traditional GEI. We do this using a new gait database which was recorded with the Kinect sensor. On this database we show significant performance gain of DGHEI.

Hypergraphs for Joint Multi-view Reconstruction and Multi-object Tracking

We generalize the network flow formulation for multiobject tracking to multi-camera setups. In the past, reconstruction of multi-camera data was done as a separate extension. In this work, we present a combined maximum a posteriori (MAP) formulation, which jointly models multicamera reconstruction as well as global temporal data association. A flow graph is constructed, which tracks objects in 3D world space. The multi-camera reconstruction can be efficiently incorporated as additional constraints on the flow graph without making the graph unnecessarily large. The final graph is efficiently solved using binary linear programming. On the PETS 2009 dataset we achieve results that significantly exceed the current state of the art.

Combined face and gait recognition using alpha matte preprocessing

This paper presents advances on the Human ID Gait Challenge. Our method is based on combining an improved gait recognition method with an adapted low resolution face recognition method. For this, we experiment with a new automated segmentation technique based on alpha-matting. This allows better construction of feature images used for gait recognition. The same segmentation is also used as a basis for finding and recognizing low-resolution facial profile images in the same database. Both, gait and face recognition methods show results comparable to the state of the art. Next, the two approaches are fused (which to our knowledge, has not yet been done for the Human ID Gait Challenge). With this fusion gain, we show significant performance improvement. Moreover, we reach the highest recognition rates and the largest absolute number of correct detections to date.

Unified hierarchical multi-object tracking using global data association

This paper presents a unified hierarchical multi-object tracking scheme. The problem of simultaneously tracking multiple objects is cast as a global MAP problem which aims at maximizing the probability of trajectories given the observations in each frame. Directly solving this problem is infeasible, due to computational considerations and the difficulty of reliably estimate necessary transition probabilities. Without breaking the MAP formulation, we propose a three stage hierarchical tracking framework which makes solving the MAP feasible. In addition, using a hierarchical framework allows for modeling inter-object occlusions. Occlusion handling thus smoothly and implicitly integrates into the proposed framework without any explicit occlusion reasoning. Finally, we evaluate the proposed method on the publicly available PETS 2009 tracking data and show improvements over the current state of the art for most sequences.

Improved Gait Recognition using Gradient Histogram Energy Image

We present a new spatio-temporal representation for Gait Recognition, which we call Gradient Histogram Energy Image (GHEI). Similar to the successful Gait Energy Image (GEI), information is averaged over full gait cycles to reduce noise. Contrary to GEI, where silhouettes are averaged and thus only edge information at the boundary is used, our GHEI computes gradient histograms at all locations of the original image. Thus, also edge information inside the person silhouette is captured. In addition, we show that GHEI can be greatly improved using precise segmentation techniques (we use α-matte segmentation). We demonstrate great effectiveness of GHEI and its variants in our experiments on the large and widely used HumanID Gait Challenge dataset. On this dataset we reach a significant performance gain over the current state of the art.

Depth gradient based segmentation of overlapping foreground objects in range images

Using standard background modeling approaches, close or overlapping objects are often detected as a single blob. In this paper we propose a new and effective method to distinguish between overlapping foreground objects in data obtained from a time of flight sensor. For this we use fusion of the infrared and the range data channels. In addition a further processing step is introduced to evaluate if connected components should be further divided. This is done using nonmaximum suppression on strong depth gradients.

Exploiting gradient histograms for gait-based person identification

In this paper, we exploit gradient histograms for person identification based on gait. A traditional and successful method for gait recognition is the Gait Energy Image (GEI). Here, person silhouettes are averaged over full gait cycles, which leads to a robust and efficient representation. However, binarized silhouettes only capture edge information at the boundary of the person. By contrast, the Gradient Histogram Energy Image (GHEI) also captures edges within the silhouette by means of gradient histograms. Combined with precise α-matte preprocessing and with a new part-based extension, recognition performance can be further improved. In addition, we show, that GEI can even be outperformed by directly applying gradient histogram extraction on the already bina-rized silhouettes. We run all experiments on the widely used HumanID gait database and show significant performance improvements over the current state of the art.

Gait-based person identification by spectral, cepstral and energy-related audio features

With this work, we address the problem of acoustic gait-based person identification, which is the task of identifying humans by the sounds they make while walking. We examine several acoustic features from speech processing tasks for their suitability for acoustic gait recognition. Using a wrapper-based feature selection technique, we reduce the feature set while at the same time increasing the identification accuracy by 10% (relative). For classification, Support Vector Machines (SVMs) are employed. Experiments are conducted using the TUM GAID database, which is a large gait recognition database containing 3 050 recordings of 305 subjects in three variations.