Junfa Liu

Cross-people mobile-phone based activity recognition

Activity recognition using mobile phones has great potential in many applications including mobile healthcare. In order to let a person easily know whether he is in strict compliance with the doctors exercise prescription and adjust his exercise amount accordingly, we can use a smart-phone based activity reporting system to accurately recognize a range of daily activities and report the duration of each activity. A triaxial accelerometer embedded in the smart phone is used for the classification of several activities, such as staying still, walking, running, and going upstairs and downstairs. The model learnt from a specific person often cannot yield accurate results when used on a different person. To solve the cross-people activity recognition problem, we propose an algorithm known as TransEMDT (Transfer learning EMbedded Decision Tree) that integrates a decision tree and the k-means clustering algorithm for personalized activity-recognition model adaptation. Tested on a real-world data set, the results show that our algorithm outperforms several traditional baseline algorithms.

SELM: Semi-supervised ELM with application in sparse calibrated location estimation

Abstract Indoor location estimation based on Wi-Fi has attracted more and more attention from both research and industry fields. It brings two significant challenges. One is requiring a vast amount of labeled calibration data. The other is real-time training and testing for location estimation task. Traditional machine learning methods cannot get high performance in both aspects. This paper proposed a novel semi-supervised learning method SELM (semi-supervised extreme learning machine) and applied it to sparse calibrated location estimation. There are two advantages of the proposed SELM. First, it employs graph Laplacian regularization to import large number of unlabeled samples which can dramatically reduce labeled calibration samples. Second, it inherits the good property of ELM on extreme training and testing speed. Comparative experiments show that with same number of labeled samples, our method outperforms original ELM and back propagation (BP) network, especially in the case that the calibration data is very sparse.

Mapping learning in eigenspace for harmonious caricature generation

This paper proposes a mapping learning approach for caricature auto-generation. Simulating the artists creativity based on the objects facial feature, our approach targets discovering what are the principal components of the facial features, and whats the difference between facial photograph and caricature measured by those components. In training phase, PCA approach is adopted to obtain the principal components. Then, machine learning of SVR (Support Vector Regression) is carried out to learn the mapping model in principal component space. With the mapping model, in application phase, users just need to input a frontal facial photograph for the caricature generation. The caricature is exaggerated based on the original face while reserving essential similar features. Experiments proved comparatively that our approach could generate more harmonious caricatures.

TOSELM: Timeliness Online Sequential Extreme Learning Machine

For handling data and training model, existing machine learning methods do not take timeliness problem into consideration. Timeliness here means the data distribution or the data trend changes with time passing by. Based on timeliness management scheme, a novel machine learning algorithm Timeliness Online Sequential Extreme Learning Machine (TOSELM) is proposed, which improves Online Sequential Extreme Learning Machine (OSELM) with central tendency and dispersion characteristics of data to deal with timeliness problem. The performance of proposed algorithm has been validated on several simulated and realistic datasets, and experimental results show that TOSELM utilizing adaptive weight scheme and iteration scheme can achieve higher learning accuracy, faster convergence and better stability than other machine learning methods.

Inferring social contextual behavior from bluetooth traces

Context-aware computing is increasingly paid much attention, especially makes the peoples social contextual behavior very crucial for user-centric dynamic behavior inference. At present, extensive work has focused on detecting specific places inferred by static radio signals like GPS, GSM and WiFi, and recognizing mobility modes inferred by embedded sensor components like accelerometer. This paper proposes a distinct feature based classification approach and context restraint based majority vote rule to infer social contextual behavior in dynamic surroundings. Experimental results indicate that our proposed method can achieve high accuracy for inferring social contextual behavior through the real-life Bluetooth traces.

Semi-supervised Learning of Caricature Pattern from Manifold Regularization

Automatic caricature synthesis is to transform the input face to an exaggerated one. It is becoming an interesting research topic, but it remains an open issue to specify the caricatures pattern for the input face. This paper proposed a novel pattern prediction method based on MR (manifold regularization), which comprises three steps. Firstly, we learn the caricature pattern by manifold dimension reduction, and select some low dimensional caricature pattern as the labels for corresponsive true faces. Secondly, manifold regularization is performed to build a semi-supervised regression between true faces and the pattern labels. In the third step of offline phase, the input face is mapped to a pattern label by the learnt regressive model, and the pattern label is further transformed to caricature parameters by a locally linear reconstruction algorithm. This approach takes advantage of manifold structure lying in both true faces and caricatures. Experiments show that, low dimensional manifold represents the caricature pattern well and the semi-supervised regressive model from manifold regularization can predict the target caricature pattern successfully.

Semi‐Supervised Learning in Reconstructed Manifold Space for 3D Caricature Generation

Recently, automatic 3D caricature generation has attracted much attention from both the research community and the game industry. Machine learning has been proven effective in the automatic generation of caricatures. However, the lack of 3D caricature samples makes it challenging to train a good model. This paper addresses this problem by two steps. First, the training set is enlarged by reconstructing 3D caricatures. We reconstruct 3D caricatures based on some 2D caricature samples with a Principal Component Analysis (PCA)‐based method. Secondly, between the 2D real faces and the enlarged 3D caricatures, a regressive model is learnt by the semi‐supervised manifold regularization (MR) method. We then predict 3D caricatures for 2D real faces with the learnt model. The experiments show that our novel approach synthesizes the 3D caricature more effectively than traditional methods. Moreover, our system has been applied successfully in a massive multi‐user educational game to provide human‐like avatars.

Fall Detecting and Alarming Based on Mobile Phone

This paper presents a system for fall detecting using off-the-shelf electronic devices to detect the fall. We use a smart phone with an embedded tri-axial accelerometer sensor. Data from the accelerometer is evaluated with a decision tree model to determine a fall. If a fall is suspected, a notification is raised to require the user’s response. If the user hurts hardly and cannot respond, the system alerts pre-specified guardian with a message via SMS. Therefore, the fallen man can be cared immediately.

Surrounding context and episode awareness using dynamic Bluetooth data

Bluetooth information can efficiently capture characteristics of user-centric surrounding contexts, such as formal meeting or chatting with friends, shopping with friends or alone, etc. In this paper, we extract novel features from Bluetooth traces and use these features for recognizing contextual behavior as well as inferring continuous episode transition. Evaluation results show that extracted novel features are very effective, which enable the model to achieve an average of 87% accuracy for specific context classification and the ability of episode inference from real-life Bluetooth traces.

Interactive 3D caricature generation based on double sampling

Recently, 3D caricature generation and applications have attracted wide attention from both the research community and the entertainment industry. This paper proposes a novel interactive approach for various and interesting 3D caricature generation based on double sampling. Firstly, according to users operation, we obtain a coarse 3D caricature with local features transformation by sampling in well-built principle component analysis (PCA) subspace. Secondly, to utilize information of the 2D caricature dataset, we sample in the local linear embedding (LLE) manifold subspace. Finally, we use the learned 2D caricature information to further refine the coarse caricature by applying Kriging interpolation. The experiments show that the 3D caricature generated by our method can preserve highly artistic styles and also reflect the users intention.