An art installation was on display in the Centre Pompidou National Museum of Modern Art in Paris, were visitors could contribute with their own personal objects, adding keyword descriptions and quantified semantic features such as age or hardness. The data was projected in real-time onto a Self-Organizing Map (SOM) and shown in the gallery. In this paper we analyze the same data by extracting v...

The distribution of a class of objects, such as images depicting a specific topic, can be studied by observing the best-matching units (BMUs) of the objects’ feature vectors on a Self-Organizing Map (SOM). When the BMU “hits” on the map are summed up, the class distribution may be seen as a two-dimensional histogram or discrete probability density. Due to the SOM’s topology preserving property,...

An adaptive two-stage approach is proposed for the classiication of surface defects. In the segmentation phase feature extraction is done and potential defect areas are marked. In the classiication phase features describing the shape and internal structure of defects are extracted and defects are classiied to diierent defect classes. The proposed segmentation scheme is taught with examples of f...

A large body of human image processing techniques use skin detection as a first primitive for subsequent feature extraction. Well established methods of colour modelling, such as histograms and Gaussian mixture models have enabled the construction of suitably accurate skin detectors. However such techniques are not ideal for use in adaptive real time environments. We describe methods of skin de...

In this paper a novel framework is presented for interactive feature-based retrieval and visualization of human statues, using depth sensors for mobile devices. A skeletal model is fitted to the depth image of a statue or human body in general and is used as a feature vector that captures the pose variations in a given collection of skeleton data. A scaleand twistinvariant distance function is ...

In this paper we study how the Self-Organizing Map (SOM) can be used in analysing the structure of semantic concepts in visual data. We investigate two data sets with concept labels provided by humans, and unlabelled data for which we utilise automatically detected concept membership scores by using models trained on a labelled data set. By arranging the concept memberships of visual objects as...

We present a system for content-based retrieval of perceptually similar sound events in audio documents (‘sound spotting’, using a query by example. The system consists of three discrete stages: a front-end for feature extraction, a self-organizing map, and a pattern matching unit. Our paper introduces the approach, describes the separate modules and discusses some preliminary results and futur...

Texture analysis has a wide range of real-world applications. This paper presents a novel technique for texture feature extraction and compares its performance with a number of other existing techniques using a benchmark image database. The proposed feature extraction technique uses 2 D D R transform and self-organizing map (SOM). A combination of 2D-DFT and SOM with optimal parameter settings ...

| We present a system for automatic mapping of complex gray-scale images onto each other. The system includes a Neocognitron-like structure for hierarchical feature extraction, a 3D Self Organising Map to determine feature classes for unsupervised training, and algorithmic methods for landmark correspondence and image warping. We present results showing successful registration of MRI brain scan...