Humans find seeing things effortless and this hides the fact that making sense of the visual world is a very difficult problem. Vision is difficult because each image we see could have been made by an infinite number of object and lighting combinations. Think of the simplest image property - grey level. The grey level of each pixel in an image is determined by the amount of light falling onto each object and the amount of light that is reflected back from each object. Dark objects have lower grey levels than light ones but even light objects have low grey levels when in shadow. We cannot tell whether we are looking at a dark object in bright light or a light object in shadow just by measuring grey levels. Even worse, when grey levels are different in different parts of an image we cannot tell if this difference is due to there being two objects or a change in the amount of light. Despite this problem humans are very good at working out the reasons for grey level changes; we CAN tell objects from shadows.One reason why we are so good at working out what's going on in images maybe that we use other properties such as colour and pattern to tell us what the grey levels mean. This idea has led to the concept of 'intrinsic images'. An intrinsic image is an image that describes one property of the scene. So instead of having a single image that mixes up shadows and object reflectances we might produce two intrinsic images one each for shadows and reflectance. Scientists have already succeeded in producing intrinsic images like these by using colour changes to work out what the grey levels mean. But, there is more than one type of shadow and more than one type of reflection. We want to improve on the existing methods by producing four intrinsic images instead of two. Our first intrinsic image will contain the type of gentle shading that is found on undulating surfaces. Our second intrinsic image will contain the hard shadows that are produced when an object blocks the light. Our third intrinsic image will describe the reflectance of matte objects and our forth image the reflections from shiny objects. To separate out these four images we will need to use additional information beyond colour. We think that surface patterns (e.g. wood grain) will provide the necessary information.Extracting four intrinsic images will be very helpful to those engineers who try to make computers understand what's going on in an image. To take just one example, humans seem to be very good at is estimating the shape of undulations on a surface from the way that it is shaded. We are so good at this that we do it automatically and the people who write computer software can trick us into thinking that their 'buttons' stand out from the screen just by adding a some highlights and shading to the edges. There are many computer programs that try to interpret shape-from-shading. While many of these programs work well they tend to assume that all changes in grey level are due to shading which is in tern due to surface undulations. We know that this assumption is not true in real pictures and these programs tend to do badly when looking at such images. But if we can produce shading only images from real images then these programs may work better.To test our ideas and decide on the best way achieve our desired results we will collect a large number of photographs of objects whose shape we either already know or can work out. We will calibrate these pictures very carefully and then use them to workout what information is conveyed by colour and pattern that can help us to workout the meaning of each grey level change. We will also test humans to see which cues they might be using. We will make our images available on the Internet so that others can try out their ideas too. We intend to work with a software company who will take the best of our ideas and implement them in a computer program that can automatically design embossed jewellery from photographs.

The use of computers to aid the manufacturing process has been long established. In particular, the use of computer numerically controlled machines to progressively remove material from a solid block to produce a finished component is now an integral part of the component industry. The hardware, software and methodologies for material removal have all matured into state of the art software packages and multi-axis machining centres. Despite this, there remains issues with machining high precision components that should require little or no finishing. There are many parameters involved in the machining process and even when optimised the physical part may exhibit unexpected errors. These may be due to a number of effects (ignoring such issues as tool wear and vibration) including: simplifications made in the computer aided manufacturing software regarding the model of the cutting tool; the need to discretise the tool to send to machine tool controller; the need for the controller to re-interpolate the required tool path and the need to control the tool to follow the (newly re-interpolated) path. To offset these effects, time consuming and expensive physical cutting trials are required in order to produce a high quality surface finish. An alternative would be to have the ability to accurately simulate the cutting process that predicts the true cutting conditions and reproduces the machined surface finish. It would be over ambitious to attempt to construct a simulation that is capable of modelling every aspect of the entire machining process in a single project. However, if a framework can be established and demonstrated for a manageable set of parameters, this could then be developed by others to incorporate other aspects including vibration, tool wear, etc. This project aims at providing such a framework for the realistic simulation of material removal using multi-axis machining tools and a robust, integrated framework for the modelling of tool path motions. Previous work at the University of Birmingham has considered the problem of determining what the actual machined part is going to be. This work was based on the principle of generating surface normals to a CAD model of the part. Machining is then simulated by using an exact model of the cutting tool and using this to truncate the normals. This has been used successfully to predict the small cusps that can remain during manufacture. It is proposed to extend this approach to deal with more complicated surface forms so that defects can be predicted and hence means for attenuating them investigated. Interest at the University of Bath has been in the use of geometric algebra to describe motions. Geometric algebra provides a framework in which rigid-body motions, both translation and rotations, can be handled in a common form. This allows techniques for free-form curves to be extended in a natural way to deal with free-form motions. It is proposed to investigate the use of the approach for describing the motion of the cutting tool and hence to improve means for interpolating (and re-interpolating) tool-path information.

3D models have a broad range of applications in many different areas such as engineering, biology, chemistry, medicine, entertainment and cultural heritage. Many 3D models are available from the Internet and other sources, resulting in a problem of how to effectively and efficiently find required 3D models (i.e., 3D model retrieval). Current research on 3D model retrieval mainly focuses on global rigid 3D model retrieval, and algorithms for solving this problem are not effective for non-rigid and partial 3D model retrieval. Because many 3D models of interest are non-rigid (such as humans, and mechanisms), and because it is often important to consider just parts of a 3D model (e.g. find a model with a particular connector), finding an efficient way to retrieve non-rigid and partial 3D models is a pressing and challenging problem. This project intends to develop robust and sensitive algorithms for non-rigid and partial 3D model retrieval. A typical shape-based 3D model retrieval algorithm consists of three main steps: model preprocessing, feature/shape descriptor extraction, and feature/shape indexing and matching. This project will investigate all three steps and develop new non-rigid and partial 3D model retrieval algorithms based on novel techniques from other research areas. Set-membership estimation from control theory will be introduced into model preprocessing and feature/shape descriptor extraction. New machine learning methods, such as affinity propagation, manifold learning and ranking, will be explored for extracting features/shape descriptors, and for feature/shape indexing and matching. The N-gram model from natural language processing will be adapted to feature/shape indexing and matching. Other new techniques from image processing and computer vision will be investigated regarding their effectiveness for non-rigid and partial 3D model retrieval. This project will also consider potential applications of the newly developed techniques. The 3D model retrieval algorithms will be evaluated jointly with Delcam plc with a view to commercial exploitation. A practical non-rigid and partial 3D model search engine will be developed and deployed on the Internet for public use.

With the development of 3D acquisition techniques, geometric models are more and more widely available. Model editing is an effective way to generate new geometric models from existing ones. This project aims to develop a new editing framework based on robust feature analysis, which has been largely unexplored. We propose to study two related problems: robust feature analysis suitable for surface editing, and improving editing in both efficiency and quality with the guidance of features. Methods developed by the project will enable more effective editing to be carried out on triangular meshes, which has potential to be useful in wide application areas. The most direct applications are in Computer Graphics, where triangular meshes are widely used for rendering, animation and arts and Computer Aided Design, where such meshes are used to represent a variety of natural and designed shapes. Robust feature analysis techniques are fundamental to more general settings when geometric models are used, so the framework focussing on feature analysis and interactive techniques may also stimulate further research.

Shape-from-shading (SFS) is a classic problem in computer vision. It aims to estimate 3D surface shape from the variations in shading in a single photographic image. The fact that it recovers shape using only a single image makes SFS attractive to a wide range of applications, especially when other 3D imaging techniques such as stereo or depth scanners are difficult to apply. Example applications can be found in topography analysis of SAR (synthetic aperture radar) images, reconstruction of medical images, inspection of microelectronics, CAD systems, and the entertainment industry. However, despite over four decades research, SFS still remains a challenging problem which is underused in real world problems, due to a lack of robustness, and sometimes implausible results. A good solution is pressing and challenging. This project intends to develop a robust and practical SFS algorithm for accurate shape recovery from real-world images. The reasons for SFS's current poor performance on real-world images have several underlying causes. The first is that the classic assumptions of orthographic projection, Lambertian reflection, and simple lighting models are inaccurate for real-world surfaces. The second reason is that SFS is an underconstrained problem: the human visual system recovers shape not only from shading, but also from outlines, shadows, and prior experience. In computer vision, little work has considered the combination of shape from shading with other visual cues and human interactions. The third and largely overlooked reason is that many real surfaces are not smooth, and have detailed features. Most existing SFS algorithms only apply to images of smooth surfaces, and tend to over-smooth any features. Based on these observations, this project will integrate techniques from such areas as feature-aware image filtering, shape from line drawing, and user interaction, to achieve more accurate shape recovery from sophisticated real-world images. An interactive platform for SFS will be developed for realistic applications. The outcome of the research will be tested on various applications in CAD and computer vision: specifically, as part of the project, we will explore the applications to bas-relief generation, and face recognition.