The analysis of medical images has been woven into the fabric of the pattern analysis and machine intelligence (PAMI) community since the earliest days of these Transactions. Initially, the efforts in this area were seen as applying pattern analysis and computer vision techniques to another interesting dataset. However, over the last two to three decades, the unique nature of the problems presented within this area of study have led to the development of a new discipline in its own right. Examples of these include: the types of image information that are acquired, the fully three-dimensional image data, the nonrigid nature of object motion and deformation, and the statistical variation of both the underlying normal and abnormal ground truth. In this paper, we look at progress in the field over the last 20 years and suggest some of the challenges that remain for the years to come.

https://hal.inria.fr/inria-00615100/document

Medical image analysis: progress over two decades and the challenges ahead

https://bio3d.colorado.edu/SerialEM/serialEMJSB.pdf

Automated electron microscope tomography using robust prediction of specimen movements.

中枢神経系疾患の治療は正常細胞（ニューロン）の機能維持を目的とするが，脳血管障害のように機能障害の原因が細胞の死滅に基づくことは多い．一方，脳腫瘍の治療においては薬物療法や放射線療法といった腫瘍細胞の死滅を目標とするものが大きな位置を占める．いずれの場合にも，細胞死の機序を理解することは各種病態や治療法の理解のうえで重要である．現在のところ最も研究の進んでいる細胞死の型はアポトーシスである．そのなかで重要な位置を占めるミトコンドリアにおける反応および抗アポトーシス因子について概要を紹介する．

Neuroscience 細胞死：最近の知見

MUCKE aims to mine a large volume of images, to structure them conceptually and to use this conceptual structuring in order to improve large-scale image retrieval. The last decade witnessed important progress concerning low-level image representations. However, there are a number problems which need to be solved in order to unleash the full potential of image mining in applications. The central problem with low-level representations is the mismatch between them and the human interpretation of image content. This problem can be instantiated, for instance, by the incapability of existing descriptors to capture spatial relationships between the concepts represented or by their incapability to convey an explanation of why two images are similar in a content-based image retrieval framework. We start by assessing existing local descriptors for image classification and by proposing to use co-occurrence matrices to better capture spatial relationships in images. The main focus in MUCKE is on cleaning large scale Web image corpora and on proposing image representations which are closer to the human interpretation of images. Consequently, we introduce methods which tackle these two problems and compare results to state of the art methods. Note: some aspects of this deliverable are withheld at this time as they are pending review. Please contact the authors for a preview.

http://ieeexplore.ieee.org/iel2/4524/12820/00592460.pdf

Image Processing

http://www.cvs.rochester.edu/knill_lab/papers/35_TINS_2004.pdf

The Bayesian brain: the role of uncertainty in neural coding and computation

Model observers have a long history of success in predicting human observer performance in clinically-relevant detection tasks. New 3D image modalities provide more signal information but vastly increase the search space to be scrutinized. Here, we compared standard linear model observers (ideal observers, non-pre-whitening matched filter with eye filter, and various versions of Channelized Hotelling models) to human performance searching in 3D 1/f$^{2.8}$ filtered noise images and assessed its relationship to the more traditional location known exactly detection tasks and 2D search. We investigated two different signal types that vary in their detectability away from the point of fixation (visual periphery). We show that the influence of 3D search on human performance interacts with the signal's detectability in the visual periphery. Detection performance for signals difficult to detect in the visual periphery deteriorates greatly in 3D search but not in 3D location known exactly and 2D search. Standard model observers do not predict the interaction between 3D search and signal type. A proposed extension of the Channelized Hotelling model (foveated search model) that processes the image with reduced spatial detail away from the point of fixation, explores the image through eye movements, and scrolls across slices can successfully predict the interaction observed in humans and also the types of errors in 3D search. Together, the findings highlight the need for foveated model observers for image quality evaluation with 3D search.

Foveated Model Observers for Visual Search in 3D Medical Images

It is well known that the optimum way to perform a signal-detection or discrimination task is to compute the likelihood ratio and compare it to a threshold. Varying the threshold generates the receiver operating characteristic (ROC) curve, and the area under this curve (AUC) is a common figure of merit for task performance. AUC can be converted to a signal-to-noise ratio, often known as d a , using a well-known formula involving an error function. The ROC curve can also be determined by psychophysical studies for humans performing the same task, and again figures of merit such as AUC and d a can be derived. Since the likelihood ratio is optimal, however, the d a values for the human must necessarily be less than those for the ideal observer, and the square of the ratio of d a (human)/d a (ideal) is frequently taken as a measure of the perceptual efficiency of the human. The applicability of this efficiency measure is limited, however, since there are very few problems for which we can actually compute d a or AUC for the ideal observer. In this paper we examine some basic mathematical properties of the likelihood ratio and its logarithm. We demonstrate that there are strong constraints on the form of the probability density functions for these test statistics. In fact, if one knows, say, the density on the logarithm of the likelihood ratio under the null hypothesis, the densities of both the likelihood and the log-likelihood are fully determined under both hypotheses. Moreover, the characteristic functions and moment-generating functions for the log-likelihood under both hypotheses are specified in terms of a likelihood-generating function. From this single function one can obtain all moments of both the likelihood and the log-likelihood under both hypotheses. Moreover, AUC is expressed to an excellent approximation by a single point on the function. We illustrate these mathematical properties by considering the problem of signal detection with uncertain signal location.

Some unlikely properties of the likelihood ratio and its logarithm

Combining imaging and genetic information to predict disease presence and progression is being codified into an emerging discipline called “radiogenomics.” Optimal evaluation methodologies for radiogenomics have not been well established. We aim to develop a decision framework based on utility analysis to assess predictive models for breast cancer diagnosis. We garnered Gail risk factors, single nucleotide polymorphisms (SNPs), and mammographic features from a retrospective case-control study. We constructed three logistic regression models built on different sets of predictive features: (1) Gail, (2) Gail + Mammo, and (3) Gail + Mammo + SNP. Then we generated receiver operating characteristic (ROC) curves for three models. After we assigned utility values for each category of outcomes (true negatives, false positives, false negatives, and true positives), we pursued optimal operating points on ROC curves to achieve maximum expected utility of breast cancer diagnosis. We performed McNemar’s test based on threshold levels at optimal operating points, and found that SNPs and mammographic features played a significant role in breast cancer risk estimation. Our study comprising utility analysis and McNemar’s test provides a decision framework to evaluate predictive models in breast cancer risk estimation.

/pdf/developing-a-utility-decision-framework-to-evaluate-2ekj26jiok.pdf

Developing a utility decision framework to evaluate predictive models in breast cancer risk estimation

In a recent paper [J. Opt. Soc. Am. A 17, 206 (2000)], a modified detectability index, d'r, was proposed to accommodate correlations between internal responses in multiple-alternative forced-choice (MAFC) experiments. The derivation given in that work pertained only to two-alternative forced choice, although it was shown empirically that the result held for general MAFC tasks when the correlation between responses is constant. Here we present a rigorous derivation that shows that the d'r result generalizes to MAFC tasks in this case.

Derivation of a detectability index for correlated responses in multiple-alternative forced-choice experiments.

An ideal observer model is developed for the task of detecting blood perfusing or flowing through tissue. The ideal observer theory relies on a linear systems model that describes tissue and blood object functions and electronic noise as random processes. When aliasing is minimal, the system is characterized by a quantity similar to Noise-Equivalent Quanta used in photon imaging modalities. A simple 1-D model is used to illustrate the effect of the system and object parameters on task performance. Velocity and decorrelation are seen to be advantageous for detection. Aliasing can degrade performance. The ideal observer model provides a framework for assessing the performance of Power Doppler ultrasound systems, and may aid in their design.

/pdf/ideal-observer-model-for-detection-of-blood-perfusion-and-121knzrdzz.pdf

Craig K. Abbey

Papers

Foveated Model Observers for Visual Search in 3D Medical Images

Some unlikely properties of the likelihood ratio and its logarithm

Developing a utility decision framework to evaluate predictive models in breast cancer risk estimation

Derivation of a detectability index for correlated responses in multiple-alternative forced-choice experiments.

Ideal observer model for detection of blood perfusion and flow using ultrasound.