Institute of Information Technology Lodz University of Technology it.p.lodz.pl Technical University of Lodz
Technical Faculty University of Lodz Technical Physics, Information Technology and Applied Mathematics The Faculty employs specialists in various areas of science. Its scientific research, although oriented at application has strong theoretical background. The students of the Faculty gather profound knowledge in the area of mathematics, physics and computer science application in technology, medicine and economy as well as practical skill of computers’ application in management and industry. Fields of study Information Technology Mathematics Technical Physics
Technical Institute of Information Technology University of Lodz Institute of Information Technology concentrates on information technology applications Areas of specialization Computer Graphics and Multimedia Software Engineering and Data Analysis Computer Games and Simulation Technology Computer Networks and Tele-information Systems Information Systems and Databases Telecommunication Information Technology (in English)
Technical ECTS Label University of Lodz TUL has been awarded the ECTS Label by the European Commission and is the only university in Poland which received this prestigious certificate which acknowledges introduction of the standards that ensure good quality of studies and international exchange.
Technical Computer vision projects University of Lodz Medical visualization and tracking system for neuronavigation purposes, M. Półrola, M. Struś, A. Wojciechowski – in cooperation with Barlicki Hospital of Lodz Neurosurgery support by means of neuronavigation control with a 3D scanner Application management by means of hand gesture analysis (invaluable in a sterile operating room) Self reflections in virtual reality, K. Guzek, M. Ostrowski, K. Toszek, M. Zippunnikowa, S. Livatino, A. Wojciechowski – in cooperation with Aalborg University Interactive self-reflections in a virtual, multi-reflection mirror cube Camera-based and avatar-based human head reflections Calculation boosting by means GPU assistance
Technical Computer vision projects University of Lodz Passive Limbs Rehabilitation Augmented Reality System, A. Wojciechowski, A. Majewski, P. Napieralski Passive limbs joints relative positions tracking Visual layer describing current and intended limbs movement range Monitoring rehabilitation advancement The read digital image using sound and the movement sensor, R.Bednarski, R.Szrajber The color conversion on sound The use of movement sensor for reading color with hands movement Prevent read color blind people
Technical Computer vision projects University of Lodz Modern technologies in the construction of a character movement, R. Bednarski, M.Pietruszka, A.Wojciechowski, P. Napieralski, D.Szajerman, K.Guzek, G.Glonek Use 3D scanning to obtain the models characters Use Motion Capture System and Vituix Omni (virtual reality interface for moving freely and naturally) to create a movement database Investigating possible to acquire parallel information of the above devices, and their integration in order to obtain a sufficiently detailed description of character movement. Study the possibility of using the data obtained in the application supporting the science of motion, for example : rehabilitation exercises, sports.
Technical Number plates identification University of Lodz Localization and recognition of car number plates in video sequences (R. Klimt, A. Tomczyk) Identification of areas where moving objects are located (regions of interest, optimization) Analysis of brightness to locate potential regions with number plates Modification of plate orientation Letter and number segmentation Machine learning and sign recognition (training data) Real-time localization and recognition (multi- threading)
Technical Sign language recognition University of Lodz Recognition of selected signs of deaf-and-dumb sign language (K. Bartyzel, P. S. Szczepaniak – in cooperation with the John Paul II Catholic University of Lublin) Linguistic-like description of object shape Modification of Brownian strings method for object localization New adaptive Kuwahara filter for noise removal without edge blurring Example-based recognition and feature extraction 85% of average success rate in controlled environment
Technical Sign language recognition University of Lodz The goal of this project was to analyze the properties of linguistic shape description. The sign language recognition was only an example application. This work can be further continued in the following directions: application of other active contour models, better feature extraction (additional data, better discrimination), optimization for real-time applications. Applicability of human gestures for remote control (P. S. Szczepaniak, K. Bartyzel, A. Tomczyk) Application for 2012 GRO program Determination of the desired remote control functionalities and the corresponding gestures Feature extraction (satisfying discrimination) Usage of active contour based methods for gesture identification Usage of Kinect for Windows sensor
Technical Sign language recognition University of Lodz InterPeter – polish sign language interpreter (G. Glonek, J. Pintera, T. Ciejka, D. Molinski, M. Jonczyk, J. Koszuk) MICROSOFT IMAGINE CUP 2010 POLISH WINNER – software design category Gaussian Mixture Model and background removal color segmentation Contour-based hand pose analysis Blob motion flow human body features extraction and body parts tracking Two way communication: sign-to-text and text-to-sign Average 72% success rate for selected subset of signs Architecture based on cloud computing (Windows Azure) Depth based and image based hand pose recognition (M. Pólrola, M. Strus, A. Wojciechowski) Classifiers and decision trees based solution Real-time (200 fps) performance Further development due to regression approach for non visible hand parts Hand’s parts position average accuracy at 78% for selected subset of hand poses
Technical Visualization of medical data University of Lodz The goal of this project was to design and implement system for visualization and measurement of teeth on the basis of X-ray images. The input to the system are sequences of images representing 2D slices of jaws. On their basis 3D image is constructed (interpolated) which can be manipulated in 3D space using any rotations, scaling or shifts. As a result 2D image is obtained which is a projection of 3D image on the observer's plane. Additional functionality of the system include advanced and computer aided measurement of surface areas or volumes of selected teeth structures in three dimensions. Fast and parallel methods of interpolation of 3D images Fast algorithms of manipulation of 3D images including rotations, shifts and scaling Segmentation of 3D images Precise mappings of cursors positions into 3D space including real dimensions of visualized objects
Technical Fast linear transforms for signal processing University of Lodz Neural Network in Fast Adaptive Fourier Descriptor Based Leaves Classication (D. Puchala, M. Yatsymirskyy) Fast Parametrized Biorthogonal Transforms with Normalized Basis Vectors (D. Puchala, M. Yatsymirskyy) Image compression and image filtering using fast orthogonal and fast parametric transforms Processing and analysis of images with aid of neural networks Image encryption based on fast parametric transformations Development of algorithms for adaptation of parametric transforms Fast biorthogonal transforms for signal filtering and data classification Image analysis with aid of fast orthogonal and unitary transformations
ld.param.u32 %buf, [d_buf]; ld.param.u32 %prm, [d_prm]; ld.param.u32 %tsc, [d_tsc]; mov.u32 %sbf, sbuf; mov.u32 %ssc, stsc; // calculate id mov.u32 %id, %tid.x; shl.b32 %id, %id, 3; // load permutations add.u32 %prm, %prm, %id; ld.global.u32 %s1, [%prm+0]; ld.global.u32 %s2, [%prm+4]; // permute shl.b32 %i1, %id, 1; add.u32 %i2, %i1, 8; add.u32 %i1, %i1, %buf; ld.global.b64 %b1, [%i1]; add.u32 %i2, %i2, %buf; Technical ld.global.b64 %b2, [%i2]; Mass-parallel computations on GPUs shl.b32 %s1, %s1, 3; shl.b32 %s2, %s2, 3; University add.u32 %s1, %s1, %sbf; st.shared.b64 [%s1], %b1; of Lodz add.u32 %s2, %s2, %sbf; st.shared.b64 [%s2], %b2; Effective implementations of Fast Fourier Transform on GPU and CPU (D. Puchala, K. Stokfiszewski, B. Szczepaniak, M. Yatsymirskyy) Effective realizations of biorthogonal wavelet transforms with lattice structures on GPU and CPU (D. Puchala, B. Szczepaniak, M. Yatsymirskyy) General purpose calculations on GPU Mass-parallel variants of algorithms for calculation of Fourier transform Mass-parallel computations of wavelet transformation Extension of designed algorithms to the cases of multidimensional data Analysis of efficiency of parallel algorithms Hardware oriented optimizations IENA 2015
Technical 3D scene modelling University of Lodz Sonification of 3D environments using scene segmentation and personalized audio (M. Bujacz, P. Skulimowski, P. Strumiłło, G. Wróblewski, A. Wojciechowski, D. Szajerman) – Faculty of EECCE of the TUL co-realized by the IT Institute staff.
Recommend
More recommend
