2009:Query-by-Singing/Humming Results

Introduction

These are the results for the 2008 running of the Query-by-Singing/Humming task. For background information about this task set please refer to the Query by Singing/Humming page.

Task Descriptions

Task 1 Goto Task 1 Results: The first subtask is the same as last year. In this subtask, submitted systems take a sung query as input and return a list of songs from the test database. Mean reciprocal rank (MRR) of the ground truth, as well as the simple hit(1)/miss(0) counting, is calculated over the top 10 returns. Two data sets are used:

• #Task 1, Jang's dataset Results Roger Jang's MIR-QBSH corpus with 48 songs as ground truth + 2000 Essen Collection MIDI noise files. See ESAC Data Homepage for more information about the Essen Collection. The queries consists of 4431 humming. All queries are from the beginning of references
• #Task 1, ThinkIT's dataset Results IOACAS corpus 1 data set with 106 songs as ground truth + 2000 Essen Collection MIDI noise files. See ESAC Data Homepage for more information about the Essen Collection. The queries consists of 355 humming. There are no "singing from beginning" gurantee.

Task 2 Goto Task 2 Results: The second subtask is the query against other humming. In the second subtask, Roger Jang's MIR-QBSH corpus has been divided into two groups (2040 as queries and 2391 as database). The query is performed against the other humming database and the top 10 closed are returned. The score is simple count of how many returns belong to the same ground truth song.

General Legend

Team ID

CSJ1 = Chun-Ta Chen and Jyh-Shing Roger Jang matched by beginning of the sond

CSJ2 = Chun-Ta Chen and Jyh-Shing Roger Jang matched by anywhere of the song

HAFR = Pierre Hanna, Julien Allali, Pascal Ferraro and Matthias Robine

Task 1, Jang's dataset Results

Task 1 Overall Results

file /nema-raid/www/mirex/results/qbsh/qbsh_task1_summary.csv not found

Task 1 Friedman's Test for Significant Differences

The Friedman test was run in MATLAB against the QBSH Task 1 MRR data over the 48 ground truth song groups. Command: [c,m,h,gnames] = multcompare(stats, 'ctype', 'tukey-kramer','estimate', 'friedman', 'alpha', 0.05); file /nema-raid/www/mirex/results/qbsh/qbsh.task1.friedman_detailed.csv not found

Task 1 Summary Results by Query Group

file /nema-raid/www/mirex/results/qbsh/qbsh.task1.res.byQueryGroup.csv not found

Task 2 Results

In this subtask, the same setup as the first subtask used with combination of different transcribers and matchers. The test databases consists of 106 ground-truth MIDIS + 2000 Essen Collection MIDI noise files. The query databases consists of 355 sung queries.

Task 2 Overall Results

file /nema-raid/www/mirex/results/qbsh/qbsh.task2.summary.csv not found

Task 2 Friedman's Test for Significant Differences

The Friedman test was run in MATLAB against the QBSH Task 1 MRR data over the 48 ground truth song groups. Command: [c,m,h,gnames] = multcompare(stats, 'ctype', 'tukey-kramer','estimate', 'friedman', 'alpha', 0.05); file /nema-raid/www/mirex/results/qbsh/qbsh.task2.friedman_detailed.csv not found

Task 2 Summary Results by Query Group

file /nema-raid/www/mirex/results/qbsh/qbsh_task2_res_byQueryGroup.csv not found

Runtime Results

file /nema-raid/www/mirex/results/qbsh.runtime.csv not found

Introduction

Team ID

CSJ = Chun-Ta Chen and Jyh-Shing Roger Jang

HAFR1 = Pierre Hanna, Julien Allali, Pascal Ferraro and Matthias Robine

HAFR2 = Pierre Hanna, Julien Allali, Pascal Ferraro and Matthias Robine

HL = Shu-Jen Show and Hsiao Tyne Liang