Learning Grasp Affordances from Vision

Monimutkaisissa ja muuttuvissa ympäristöissä työskentelevät robotit tarvitsevat kykyä manipuloida ja tarttua esineisiin. Tämä työ tutkii robottitarttumisen ja robottitartuntapis-teiden koneoppimisen aiempaa tutkimusta ja nykytilaa. Nykyaikaiset menetelmät käydään läpi, ja Le:n koneoppimiseen pohjautuva luokitin toteutetaan, koska se tarjoaa parhaan onnistumisprosentin tutkituista menetelmistä ja on muokattavissa sopivaksi käytettävissä olevalle robotille. Toteutettu menetelmä käyttää intensititeettikuvaan ja syvyyskuvaan po-hjautuvia ominaisuuksi luokitellakseen potentiaaliset tartuntapisteet. Tämän toteutuksen tulokset esitellään.

Probabilistic position and orientation estimation of rigid objects using an RGB-D sensor

The recent emergence of low-cost RGB-D sensors has brought new opportunities for robotics by providing affordable devices that can provide synchronized images with both color and depth information. In this thesis, recent work on pose estimation utilizing RGBD sensors is reviewed. Also, a pose recognition system for rigid objects using RGB-D data is implemented. The implementation uses half-edge primitives extracted from the RGB-D images for pose estimation. The system is based on the probabilistic object representation framework by Detry et al., which utilizes Nonparametric Belief Propagation for pose inference. Experiments are performed on household objects to evaluate the performance and robustness of the system.

Feature-based camera pose estimation

The estimating of the relative orientation and position of a camera is one of the integral topics in the field of computer vision. The accuracy of a certain Finnish technology company’s traffic sign inventory and localization process can be improved by utilizing the aforementioned concept. The company’s localization process uses video data produced by a vehicle installed camera. The accuracy of estimated traffic sign locations depends on the relative orientation between the camera and the vehicle. This thesis proposes a computer vision based software solution which can estimate a camera’s orientation relative to the movement direction of the vehicle by utilizing video data. The task was solved by using feature-based methods and open source software. When using simulated data sets, the camera orientation estimates had an absolute error of 0.31 degrees on average. The software solution can be integrated to be a part of the traffic sign localization pipeline of the company in question.

Deep generative models for facial keypoints detection

A new area of machine learning research called deep learning, has moved machine learning closer to one of its original goals: artificial intelligence and general learning algorithm. The key idea is to pretrain models in completely unsupervised way and finally they can be fine-tuned for the task at hand using supervised learning. In this thesis, a general introduction to deep learning models and algorithms are given and these methods are applied to facial keypoints detection. The task is to predict the positions of 15 keypoints on grayscale face images. Each predicted keypoint is specified by an (x,y) real-valued pair in the space of pixel indices. In experiments, we pretrained deep belief networks (DBN) and finally performed a discriminative fine-tuning. We varied the depth and size of an architecture. We tested both deterministic and sampled hidden activations and the effect of additional unlabeled data on pretraining. The experimental results show that our model provides better results than publicly available benchmarks for the dataset.

Supervised localimage feature detection

This thesis is about detection of local image features. The research topic belongs to the wider area of object detection, which is a machine vision and pattern recognition problem where an object must be detected (located) in an image. State-of-the-art object detection methods often divide the problem into separate interest point detection and local image description steps, but in this thesis a different technique is used, leading to higher quality image features which enable more precise localization. Instead of using interest point detection the landmark positions are marked manually. Therefore, the quality of the image features is not limited by the interest point detection phase and the learning of image features is simplified. The approach combines both interest point detection and local description into one phase for detection. Computational efficiency of the descriptor is therefore important, leaving out many of the commonly used descriptors as unsuitably heavy. Multiresolution Gabor features has been the main descriptor in this thesis and improving their efficiency is a significant part. Actual image features are formed from descriptors by using a classifierwhich can then recognize similar looking patches in new images. The main classifier is based on Gaussian mixture models. Classifiers are used in one-class classifier configuration where there are only positive training samples without explicit background class. The local image feature detection method has been tested with two freely available face detection databases and a proprietary license plate database. The localization performance was very good in these experiments. Other applications applying the same under-lying techniques are also presented, including object categorization and fault detection.

New Distance Transforms for Gray Level Image Compression

This thesis deals with distance transforms which are a fundamental issue in image processing and computer vision. In this thesis, two new distance transforms for gray level images are presented. As a new application for distance transforms, they are applied to gray level image compression. The new distance transforms are both new extensions of the well known distance transform algorithm developed by Rosenfeld, Pfaltz and Lay. With some modification their algorithm which calculates a distance transform on binary images with a chosen kernel has been made to calculate a chessboard like distance transform with integer numbers (DTOCS) and a real value distance transform (EDTOCS) on gray level images. Both distance transforms, the DTOCS and EDTOCS, require only two passes over the graylevel image and are extremely simple to implement. Only two image buffers are needed: The original gray level image and the binary image which defines the region(s) of calculation. No other image buffers are needed even if more than one iteration round is performed. For large neighborhoods and complicated images the two pass distance algorithm has to be applied to the image more than once, typically 3 10 times. Different types of kernels can be adopted. It is important to notice that no other existing transform calculates the same kind of distance map as the DTOCS. All the other gray weighted distance function, GRAYMAT etc. algorithms find the minimum path joining two points by the smallest sum of gray levels or weighting the distance values directly by the gray levels in some manner. The DTOCS does not weight them that way. The DTOCS gives a weighted version of the chessboard distance map. The weights are not constant, but gray value differences of the original image. The difference between the DTOCS map and other distance transforms for gray level images is shown. The difference between the DTOCS and EDTOCS is that the EDTOCS calculates these gray level differences in a different way. It propagates local Euclidean distances inside a kernel. Analytical derivations of some results concerning the DTOCS and the EDTOCS are presented. Commonly distance transforms are used for feature extraction in pattern recognition and learning. Their use in image compression is very rare. This thesis introduces a new application area for distance transforms. Three new image compression algorithms based on the DTOCS and one based on the EDTOCS are presented. Control points, i.e. points that are considered fundamental for the reconstruction of the image, are selected from the gray level image using the DTOCS and the EDTOCS. The first group of methods select the maximas of the distance image to new control points and the second group of methods compare the DTOCS distance to binary image chessboard distance. The effect of applying threshold masks of different sizes along the threshold boundaries is studied. The time complexity of the compression algorithms is analyzed both analytically and experimentally. It is shown that the time complexity of the algorithms is independent of the number of control points, i.e. the compression ratio. Also a new morphological image decompression scheme is presented, the 8 kernels' method. Several decompressed images are presented. The best results are obtained using the Delaunay triangulation. The obtained image quality equals that of the DCT images with a 4 x 4

Post-processing and analysis of tracked hand trajectories

In this thesis, the suitability of different trackers for finger tracking in high-speed videos was studied. Tracked finger trajectories from the videos were post-processed and analysed using various filtering and smoothing methods. Position derivatives of the trajectories, speed and acceleration were extracted for the purposes of hand motion analysis. Overall, two methods, Kernelized Correlation Filters and Spatio-Temporal Context Learning tracking, performed better than the others in the tests. Both achieved high accuracy for the selected high-speed videos and also allowed real-time processing, being able to process over 500 frames per second. In addition, the results showed that different filtering methods can be applied to produce more appropriate velocity and acceleration curves calculated from the tracking data. Local Regression filtering and Unscented Kalman Smoother gave the best results in the tests. Furthermore, the results show that tracking and filtering methods are suitable for high-speed hand-tracking and trajectory-data post-processing.

Generative Part-Based Gabor Object Detector

Object detection is a fundamental task of computer vision that is utilized as a core part in a number of industrial and scientific applications, for example, in robotics, where objects need to be correctly detected and localized prior to being grasped and manipulated. Existing object detectors vary in (i) the amount of supervision they need for training, (ii) the type of a learning method adopted (generative or discriminative) and (iii) the amount of spatial information used in the object model (model-free, using no spatial information in the object model, or model-based, with the explicit spatial model of an object). Although some existing methods report good performance in the detection of certain objects, the results tend to be application specific and no universal method has been found that clearly outperforms all others in all areas. This work proposes a novel generative part-based object detector. The generative learning procedure of the developed method allows learning from positive examples only. The detector is based on finding semantically meaningful parts of the object (i.e. a part detector) that can provide additional information to object location, for example, pose. The object class model, i.e. the appearance of the object parts and their spatial variance, constellation, is explicitly modelled in a fully probabilistic manner. The appearance is based on bio-inspired complex-valued Gabor features that are transformed to part probabilities by an unsupervised Gaussian Mixture Model (GMM). The proposed novel randomized GMM enables learning from only a few training examples. The probabilistic spatial model of the part configurations is constructed with a mixture of 2D Gaussians. The appearance of the parts of the object is learned in an object canonical space that removes geometric variations from the part appearance model. Robustness to pose variations is achieved by object pose quantization, which is more efficient than previously used scale and orientation shifts in the Gabor feature space. Performance of the resulting generative object detector is characterized by high recall with low precision, i.e. the generative detector produces large number of false positive detections. Thus a discriminative classifier is used to prune false positive candidate detections produced by the generative detector improving its precision while keeping high recall. Using only a small number of positive examples, the developed object detector performs comparably to state-of-the-art discriminative methods.

Deep convolutional neural networks for semantic video object segmentation

In this thesis, we propose to infer pixel-level labelling in video by utilising only object category information, exploiting the intrinsic structure of video data. Our motivation is the observation that image-level labels are much more easily to be acquired than pixel-level labels, and it is natural to find a link between the image level recognition and pixel level classification in video data, which would transfer learned recognition models from one domain to the other one. To this end, this thesis proposes two domain adaptation approaches to adapt the deep convolutional neural network (CNN) image recognition model trained from labelled image data to the target domain exploiting both semantic evidence learned from CNN, and the intrinsic structures of unlabelled video data. Our proposed approaches explicitly model and compensate for the domain adaptation from the source domain to the target domain which in turn underpins a robust semantic object segmentation method for natural videos. We demonstrate the superior performance of our methods by presenting extensive evaluations on challenging datasets comparing with the state-of-the-art methods.

Katsekulmien vaikutus näönrasitusoireisiin näyttöpäätetyössä

Päätetyöhön epäillään liittyvän monenlaisia ongelmia. Eniten epäiltyjä ja käsiteltyjä ovat silmien rasitus- ja ärsytysoireet sekä päätetyön kuormittavuus ja näköergonomiset ongelmat. Näkemiseen ja silmiin liittyvät ongelmat näyttöpäätetyöskentelyssä ovat hyvin tavallisia. Niitä kutsutaan termillä Computer Vision Syndrome (CVS). Opinnäytetyömme tarkoituksena oli tutkia kuinka eri katsekulmat vaikuttavat näönrasitusoireisiin sekä olemassa oleviin näköjärjestelmän vikoihin. Kokeessa näyttöpääte sijoitettiin kolmeen eri katsekulmaan. Nämä kulmat olivat 15 astetta horisontaalilinjan yläpuolelle, horisontaalilinja sekä 15 astetta horisontaalilinjan alapuolelle. Tutkimus oli vertaileva ikäryhmien 20-39 ja 40-60-vuotiaat välillä. Opinnäytetyö on kvantitatiivinen. Tutkimusjoukko koostui 80 henkilöstä. VSQ- ja SSQ-kyselylomakkeilla ja mittauksilla saatu aineisto analysoitiin SPSS-ohjelmassa Wilcoxonin merkkitestillä ja Mann-Whitneyn U-testillä. Koko tutkimusjoukon SSQ-oireiden keskiarvoja tarkastellessa voitiin oireiden todeta voimistuneen tehtävän aikana tilastollisesti merkitsevästi. + 15 asteen katsekulmassa havaittiin oireiden voimistumista eniten. SSQ-oireiden jakaminen eri ryhmiin toi esiin tilastollisesti merkitseviä eroja varsinkin silmänrasitusoireiden kohdalla. - 15 asteen katsekulma aiheutti vähiten oireiden arvojen kasvua tehtävän aikana silmänrasitus- ja disorientaatio-oireiden ryhmissä. Tarkasteltaessa koko joukon silmänrasitus- ja disorientaatio-oireita voidaan päätellä näyttöpäätetyön aiheuttavan rasitusoireiden lisääntymistä, koska merkitsevyystaso näissä oli tilastollisesti erittäin merkitsevä. Sekä kokonaisuudessaan että oireryhmittäin oli huomionarvoista, että 20-40-vuotiaat kokivat näyttöpäätetyön rasittavan enemmän. Mittaustulosten perusteella voidaan sanoa, että akkommodaatiolaajuus ja konvergenssikyky olivat merkitsevästi heikompia tehtävän jälkeen. Kyynelfilmin repeämisajan keskiarvo kokeen jälkeen koko tutkimusjoukolla oli normaaliarvoa alhaisempi. Yhteistyökumppanimme voi hyödyntää työmme tuloksia laajemmassa tutkimuksessa. Opinnäytetyömme tukee ammattiosaamistamme toimiessamme näönhuollon asiantuntijoina.

Real-time Imaging and Mosaicking of Planar Surfaces

Laajojen pintojen kuvaaminen rajoitetussa työskentelytilassa riittävällä kuvatarkkuudella voi olla vaikeaa. Kuvaaminen on suoritettava osissa ja osat koottava saumattomaksi kokonaisnäkymäksi eli mosaiikkikuvaksi. Kuvauslaitetta käsin siirtelevän käyttäjän on saatava välitöntä palautetta, jotta mosaiikkiin ei jäisi aukkoja ja työ olisi nopeaa. Työn tarkoituksena oli rakentaa pieni, kannettava ja tarkka kuvauslaite paperi- ja painoteollisuuden tarpeisiin sekä kehittää palautteen antamiseen menetelmä, joka koostaaja esittää karkeaa mosaiikkikuvaa tosiajassa. Työssä rakennettiin kaksi kuvauslaitetta: ensimmäinen kuluttajille ja toinen teollisuuteen tarkoitetuista osista. Kuvamateriaali käsiteltiin tavallisella pöytätietokoneella. Videokuvien välinen liike laskettiin yksinkertaisella seurantamenetelmällä ja mosaiikkikuvaa koottiin kameroiden kuvanopeudella. Laskennallista valaistuksenkorjausta tutkittiin ja kehitetty menetelmä otettiin käyttöön. Ensimmäisessä kuvauslaitteessa on ongelmia valaistuksen ja linssivääristymien kanssa tuottaen huonolaatuisia mosaiikkikuvia. Toisessa kuvauslaitteessa nämä ongelmat on korjattu. Seurantamenetelmä toimii hyvin ottaen huomioon sen yksinkertaisuuden ja siihen ehdotetaan monia parannuksia. Työn tulokset osoittavat, että tosiaikainen mosaiikkikuvan koostaminen megapikselin kuvamateriaalista on mahdollista kuluttajille tarkoitetulla tietokonelaitteistolla.

Multispectral image compression using the wavelet transform

Multispectral images contain information from several spectral wavelengths and currently multispectral images are widely used in remote sensing and they are becoming more common in the field of computer vision and in industrial applications. Typically, one multispectral image in remote sensing may occupy hundreds of megabytes of disk space and several this kind of images may be received from a single measurement. This study considers the compression of multispectral images. The lossy compression is based on the wavelet transform and we compare the suitability of different waveletfilters for the compression. A method for selecting a wavelet filter for the compression and reconstruction of multispectral images is developed. The performance of the multidimensional wavelet transform based compression is compared to other compression methods like PCA, ICA, SPIHT, and DCT/JPEG. The quality of the compression and reconstruction is measured by quantitative measures like signal-to-noise ratio. In addition, we have developed a qualitative measure, which combines the information from the spatial and spectral dimensions of a multispectral image and which also accounts for the visual quality of the bands from the multispectral images.

Local and Global Feature Extraction for Invariant Object Recognition

Perceiving the world visually is a basic act for humans, but for computers it is still an unsolved problem. The variability present innatural environments is an obstacle for effective computer vision. The goal of invariant object recognition is to recognise objects in a digital image despite variations in, for example, pose, lighting or occlusion. In this study, invariant object recognition is considered from the viewpoint of feature extraction. Thedifferences between local and global features are studied with emphasis on Hough transform and Gabor filtering based feature extraction. The methods are examined with respect to four capabilities: generality, invariance, stability, and efficiency. Invariant features are presented using both Hough transform and Gabor filtering. A modified Hough transform technique is also presented where the distortion tolerance is increased by incorporating local information. In addition, methods for decreasing the computational costs of the Hough transform employing parallel processing and local information are introduced.

Level Set Methods in Image Processing

This thesis gives an overview of the use of the level set methods in the field of image science. The similar fast marching method is discussed for comparison, also the narrow band and the particle level set methods are introduced. The level set method is a numerical scheme for representing, deforming and recovering structures in an arbitrary dimensions. It approximates and tracks the moving interfaces, dynamic curves and surfaces. The level set method does not define how and why some boundary is advancing the way it is but simply represents and tracks the boundary. The principal idea of the level set method is to represent the N dimensional boundary in the N+l dimensions. This gives the generality to represent even the complex boundaries. The level set methods can be powerful tools to represent dynamic boundaries, but they can require lot of computing power. Specially the basic level set method have considerable computational burden. This burden can be alleviated with more sophisticated versions of the level set algorithm like the narrow band level set method or with the programmable hardware implementation. Also the parallel approach can be used in suitable applications. It is concluded that these methods can be used in a quite broad range of image applications, like computer vision and graphics, scientific visualization and also to solve problems in computational physics. Level set methods and methods derived and inspired by it will be in the front line of image processing also in the future.