Tapio Manninen

The 9th annual MLSP competition: New methods for acoustic classification of multiple simultaneous bird species in a noisy environment

Birds have been widely used as biological indicators for ecological research. They respond quickly to environmental changes and can be used to infer about other organisms (e.g., insects they feed on). Traditional methods for collecting data about birds involves costly human effort. A promising alternative is acoustic monitoring. There are many advantages to recording audio of birds compared to human surveys, including increased temporal and spatial resolution and extent, applicability in remote sites, reduced observer bias, and potentially lower cost. However, it is an open problem for signal processing and machine learning to reliably identify bird sounds in real-world audio data collected in an acoustic monitoring scenario. Some of the major challenges include multiple simultaneously vocalizing birds, other sources of non-bird sound (e.g., buzzing insects), and background noise like wind, rain, and motor vehicles.

Leukemia Prediction Using Sparse Logistic Regression

We describe a supervised prediction method for diagnosis of acute myeloid leukemia (AML) from patient samples based on flow cytometry measurements. We use a data driven approach with machine learning methods to train a computational model that takes in flow cytometry measurements from a single patient and gives a confidence score of the patient being AML-positive. Our solution is based on an regularized logistic regression model that aggregates AML test statistics calculated from individual test tubes with different cell populations and fluorescent markers. The model construction is entirely data driven and no prior biological knowledge is used. The described solution scored a 100% classification accuracy in the DREAM6/FlowCAP2 Molecular Classification of Acute Myeloid Leukaemia Challenge against a golden standard consisting of 20 AML-positive and 160 healthy patients. Here we perform a more extensive validation of the prediction model performance and further improve and simplify our original method showing that statistically equal results can be obtained by using simple average marker intensities as features in the logistic regression model. In addition to the logistic regression based model, we also present other classification models and compare their performance quantitatively. The key benefit in our prediction method compared to other solutions with similar performance is that our model only uses a small fraction of the flow cytometry measurements making our solution highly economical.

Bayesian error estimation and model selection in sparse logistic regression

Regularized logistic regression models have recently become an important classification tool for high dimensional problems due to their sparseness and embedded feature selection property of the ℓ1 penalty. However, the degree of sparseness is determined by a regularization parameter λ, whose selection is typically done by cross validation. In this paper we study the applicability of a recently proposed Bayesian error estimation approach for the selection of a proper model along the regularization path. The model selection by the new Bayesian error estimator is experimentally shown to improve the classification accuracy in small sample-size situations, and is able to avoid the excess variability inherent to traditional cross validation approaches.

Object detection for dynamic adaptation of interconnections in inkjet printed electronics

A computer vision system for automatic detection of interconnection points in inkjet printed interconnected electronics modules during the manufacturing process is described. The location data can be used for calculating the amount of misalignments and the required corrections. The method uses a local filtering operation for finding candidate regions, from which the false matches are filtered out by classification using an artificial neural network. Experiments show that atypical false match rate is as low as 0.18%. The location data is later used for correcting the wiring that is inkjetted on top of the layout to connect the connectors correctly.

Point pattern matching for 2-D point sets with regular structure

Point pattern matching (PPM) is a widely studied problem in algorithm research and has numerous applications, e.g., in computer vision. In this paper we focus on a class of brute force PPM algorithms suitable for situations where the state-of-the-art methods do not perform optimally, e.g., due to point sets with regular structure. We discuss of an existing algorithm, which is optimal in the sense of brute force testing of different point pairings. We propose a parameter choosing scheme that minimizes the memory consumption of the algorithm. We also present a modified version of the algorithm to overcome issues related to its implementation and accuracy. Due to its brute force nature, the algorithm is guaranteed to return the best possible result.

Dynamic adaptation of interconnections in inkjet printed electronics

Printed electronics is a new technology for manufacturing miniature electronics modules. The basic principle is that the integrated circuits are molded into a background material such that the connectors are left visible on top. After the background substrate has hardened, the wiring is printed on top of the module using conductive ink. The technology allows flexible manufacturing of significantly smaller modules using wide range of new materials. Typically the components and their connection points are slightly displaced when the background material hardens. This paper proposes a method for adjusting the wiring to match the displaced components. Experiments show that correction reduces the number of false or missing connections significantly thus providing necessary yield improvement for the manufacturing process.

The eighth annual MLSP competition: Second place team

This paper describes our submission to the eighth annual MLSP competition organized by Amazon during the 2012 IEEE MLSP workshop. Our approach is based on a nearest-neighbor-like classifier with a distance metric learned from samples. The method was second in the final standings with prediction accuracy of 81 %, while the winning submission was 87 % accurate.

Effects of latanoprost in iris bioidentification

Purpose:  Bioidentification is becoming increasingly important in everyday life. One of the most widespread methods of bioidentification is based on the structure of the iris. Iris photography has several advantages as an identification method: it is relatively simple and effective; it is non‐invasive, and it is comparatively inexpensive. However, some medical conditions may change the appearance of the iris. This paper discusses the effects of latanoprost‐induced pigmentation changes in iris bioidentification.