Suvadip Mukherjee

Region Based Segmentation in Presence of Intensity Inhomogeneity Using Legendre Polynomials

We propose a novel region based segmentation method capable of segmenting objects in presence of significant intensity variation. Current solutions use some form of local processing to tackle intra-region inhomogeneity, which makes such methods susceptible to local minima. In this letter, we present a framework which generalizes the traditional Chan-Vese algorithm. In contrast to existing local techniques, we represent the illumination of the regions of interest in a lower dimensional subspace using a set of pre-specified basis functions. This representation enables us to accommodate heterogeneous objects, even in presence of noise. We compare our results with three state of the art techniques on a dataset focusing on biological/biomedical images with tubular or filamentous structures. Quantitatively, we achieve a 44% increase in performance, which demonstrates efficacy of the method.

Tubularity Flow Field—A Technique for Automatic Neuron Segmentation

A segmentation framework is proposed to trace neurons from confocal microscopy images. With an increasing demand for high throughput neuronal image analysis, we propose an automated scheme to perform segmentation in a variational framework. Our segmentation technique, called tubularity flow field (TuFF) performs directional regional growing guided by the direction of tubularity of the neurites. We further address the problem of sporadic signal variation in confocal microscopy by designing a local attraction force field, which is able to bridge the gaps between local neurite fragments, even in the case of complete signal loss. Segmentation is performed in an integrated fashion by incorporating the directional region growing and the attraction force-based motion in a single framework using level sets. This segmentation is accomplished without manual seed point selection; it is automated. The performance of TuFF is demonstrated over a set of 2D and 3D confocal microscopy images where we report an improvement of >75% in terms of mean absolute error over three extensively used neuron segmentation algorithms. Two novel features of the variational solution, the evolution force and the attraction force, hold promise as contributions that can be employed in a number of image analysis applications.

Tree2Tree2: Neuron tracing in 3D

We seek a complete description for the neurome of the Drosophila, which involves tracing more than 20,000 neurons. The currently available tracings are sensitive to background clutter and poor contrast of the images. In this paper, we present Tree2Tree2, an automatic neuron tracing algorithm to segment neurons from 3D confocal microscopy images. Building on our previous work in segmentation [1], this method uses an adaptive initial segmentation to detect the neuronal portions, as opposed to a global strategy that often results in under segmentation. In order to connect the disjoint portions, we use a technique called Path Search, which is based on a shortest path approach. An intelligent pruning step is also implemented to delete undesired branches. Tested on 3D confocal microscopy images of GFP labeled Drosophila neurons, the visual and quantitative results suggest that Tree2Tree2 is successful in automatically segmenting neurons in images plagued by background clutter and filament discontinuities.

Dictionary Learning Level Set

We propose a novel region based segmentation technique using dictionary learning. In a previous work we have developed a method which uses a set of pre-specified Legendre basis functions to perform region based segmentation of an object in presence of heterogeneous illumination. We hypothesize that in problems where a set of training images for the object is available for analysis (such as depth image sequence of blood vessels via ultrasound imaging), segmentation accuracy can be significantly improved by learning the basis functions instead of specifying them implicitly. The salient idea of this letter is to compute the optimal set of functions to model the region intensities. Our solution to this problem involves the integration of a level set segmentation methodology with the dictionary learning framework. This provides an elegant solution to deal with intensity inhomogeneities prevalent in many imaging applications such as ultrasound and fluorescence microscopy. The proposed algorithm, Dictionary Learning Level Set (DL2S) is used to segment ultrasound images of blood vessels captured using low cost, portable ultrasound devices employed in a phlebotomy application. Qualitative and quantitative results obtained from this dataset suggest efficacy of D2LS with an associated improvement in the average Dice index of 12% over the relevant competitors.

Visual attraction in Drosophila larvae develops during a critical period and is modulated by crowding conditions

The development of social behavior is poorly understood. Many animals adjust their behavior to environmental conditions based on a social context. Despite having relatively simple visual systems, Drosophila larvae are capable of identifying and are attracted to the movements of other larvae. Here, we show that Drosophila larval visual recognition is encoded by the movements of nearby larvae, experienced during a specific developmental critical period. Exposure to moving larvae, only during a specific period, is sufficient for later visual recognition of movement. Larvae exposed to wild-type body movements, during the critical period, are not attracted to the movements of tubby mutants, which have altered morphology. However, exposure to tubby, during the critical period, results in tubby recognition at the expense of wild-type recognition indicating that this is true learning. Visual recognition is not learned in excessively crowded conditions, and this is emulated by exposure, during the critical period, to food previously used by crowded larvae. We propose that Drosophila larvae have a distinct critical period, during which they assess both social and resource conditions, and that this irreversibly determines later visually guided social behavior. This model provides a platform towards understanding the regulation and development of social behavior.

A geometric-statistical approach toward neuron matching

In the same vein as the genome project that mapped the genetic structure of complex organisms such as the mouse, those pursuing the neurome are seeking a map the neural anatomy. In this massive biological investigation, the tools of imaging and biological experimentation are outpacing the requisite tools in image analysis. In terms of comparing neurons, based on the geometrical structure and features within the structure, the accepted approaches are largely manual. In this paper, we propose a combined geometric-statistical approach toward automated neuron matching. We utilize the geometric information of a neuron and compute a pairwise distance histogram based on the geometric information, to find a similarity measure between neurons. The distribution function is so chosen such that it reflects the structural pattern of a set of neuronal points, and is rotationally invariant. Preliminary experiments on a set of three different classes of neurons, with six neurons in each class, provides evidence of efficacy, with the best two matches to a given query producing a retrieval error of 0% and the third best match producing an error of only 11.2%. In future work, the proposed method can be used as a component in the retrieval of similar neurons from neuronal database.

Oriented filters for vessel contrast enhancement with local directional evidence

Vascular structures occur in abundance within biomedical and biological image processing applications. From detecting retinal blood vessels to analyzing shape and connectivity of neurons, segmentation of vascular structures has received significant attention in the literature. Robust segmentation often demands a preprocessing stage involving enhancement of the tubular objects. We propose a novel method to enhance vascular structures from low contrast images by incorporating evidence of neighboring tubular structures in addition to the local vessel detection. We show that the proposed algorithm, called local directional evidence (LDE), is capable of handling bifurcations, intensity inhomogeneity and complex geometry of the vessels, thus providing a robust preprocessing for segmentation. Experiments on a collection of biological images containing vascular objects suggest efficacy of LDE when used as a precursor to segmentation. We observe that LDE improves the average segmentation performance by 63% on our database over the vessel enhancing filter of [1].

Vector field convolution medialness applied to neuron tracing

In this paper we propose a novel approach to the extraction of medial axis for grayscale objects. The method utilizes a computationally efficient vector field convolution to enhance the medialness feature. Local maxima of medialness are analyzed in scale space, yielding a robust medial axis for grayscale imagery. An important application of this work is the segmentation of neurons from noisy, cluttered microscopy images. Existing neuron segmentation methods depend heavily on accurate, noise-insensitive medial axis extraction. We propose the vector field convolution medialness operation as a first step in segmenting neurons. The proposed method requires no complex parameters or an initial binarization step. The efficacy of the method is demonstrated by a 60% reduction root mean squared error (2.9 pixels) as compared to an approach based on gradient vector flow.

Shape descriptors based on compressed sensing with application to neuron matching

In this paper we propose a novel compressed sensing based Fourier shape descriptor method to compute the shape feature vector of an arbitrary object. First, the object contour obtained via segmentation is represented as a complex-valued signal. We then formulate an optimization problem that exploits the sparsity of the shape feature of the contour. This results in a reduced size feature vector, which can efficiently represent the shape of an object as illustrated by the reconstruction results. Appropriate for general shape retrieval problems, we demonstrate the efficacy of our algorithm by retrieving structurally similar neurons from a database. Currently, the representation and matching of neurons, given the heterogeneous nature of the neuronal morphology and the characteristically complex branching patterns, is an open problem. Retrieval of structurally similar neurons will potentially enable classification of neurons imaged. The retrieval results obtained using our method provide evidence of efficacy with a 27% improvement over Sholl analysis, which is a standard shape descriptor used in neuroscience.

Tracking sunflower circumnutation using affine parametric active contours

Study of the sunflower movement may reveal clues regarding unknown mechanisms that regulate periodicity and spatial complexity in plant growth and development. In this paper, we introduce an automated process to track circumnutation of sunflower seedlings. The objective is to track the leaves of the sunflower plant in a video captured by an overhead camera. The tracking method presented predicts the translated and rotated boundary in the subsequent frames by active contour models. A salient feature of our solution is a constraint on affine transformation between updates. The constrained affine active contours used in this paper exhibit improvement over other traditional active contour approaches, with the new method yielding error less than one percent in the tracked sunflower centroid position.