Professor Ben-Arie's areas of specialization are in Computer Vision, Image Processing, Neural Networks, Human Computer Interaction, Human Audition, Video and Pictorial Databases and Man-Machine Systems. In these areas he has published over 120 technical publications. So far, he has completed in these areas more than 15 research projects funded by NSF, DARPA, Whitaker Foundation, US Army Research Office (ARO), Israeli Ministry of Science, Israeli Aircraft Industry and others. The cumulative budget of these research projects exceeds $3,000,000.

Professor Ben-Arie joined UIC in August 1995 and established here the Machine Vision Laboratory (MVL). Since arriving at UIC he has received 7 research grants (6 as a PI) with the cumulative budget of $1,201,880. This total includes only his share in joint grants. The most recent grant was awarded to him as a single investigator in the framework of the prestigious NSF's program of Digital Libraries DLI2. The project is a digital library for human movement. This is a relatively new area in his research and he plans to work intensively both in digital libraries as databases for visual-temporal information and in human computer interaction (HCI). His main interest here is in developing new methods for activity analysis of humans based on computer vision techniques. He plans to incorporate his ongoing research in visual object/face recognition in image/signal processing and in neural networks in the work on the digital library. Very recently he had excellent results in human activity recognition from videos. He developed a novel recognition method of human actions that requires only sparse sampling of video streams. Unlike common methods that use HMM (Hidden Markov Models), this method employs an innovative and much more efficient scheme of parallel- indexing. The method provides not only robust recognition of activities (such as walking, running, sitting, etc.) but also an efficient database representation of all the activities.

Currently, Professor Ben-Arie has 7 graduate students and he plans to intensify his efforts in Machine Vision research. He also hopes that his current work will yield more publications. So far, he has authored 5 book chapters, 31 Journal papers, 69 refereed conferences, 3 unrefereed conferences, 1 US patent, 2 theses and 18 technical reports (average length of 62 pages each). The vast majority of his journal papers are full length papers in the top journals in his areas (such as 6 journals are in the IEEE Trans. On Pattern Analysis and Machine Intelligence - PAMI, and 5 journals in IEEE Trans. on Image Processing). Many of the refereed conference papers are in the top conferences as well (such as 8 papers in IAPR, 8 in IEEE CVPR, 8 in IEEE IP and 3 in IEEE ICASSP). He also hopes that his recent appointment as an Associate Editor to the IEEE Trans. On Image Processing will get him more involved in scientific editing as well. He also would like to combine in his current research the scientific discoveries he had during his career.

Professor Ben-Arie discovered the widely cited probabilistic peaking effect of viewed angles and distances, which is a general visual-geometric phenomenon and currently serves as the basis of the area of quasi-invariants in image understanding. He also developed the non-orthogonal expansion matching method (EXM) which is very useful in the recognition of highly occluded objects, tracking and in motion video compression. More recently he developed the Affine Invariant Spectral Signatures (AISS) and the Volumetric Frequency Representation (VFR). The AISS and the VFR provide novel, efficient means for object and face recognition and representation. The VFR is an innovative representation of objects in 3D frequency domain. He also has contributed in numerous other topics such as: Shape from Recognition, Model Based Segmentation, Neural Modeling of Auditory Localization, Evolutional design of the Human Ear, Conveying Visual Information with Spatial Auditory Patterns, Shape Description by Magnetic Fields Modeling and Pose Invariant Face Recognition using 3D-frequency domain modeling.

As mentioned above, he also developed very recently a breakthrough method for human activity recognition. This method advances the area of human video analysis to a new level. Based upon his recent achievements Dr. Ben-Arie was promoted to a full Professor in July 2001.