220 likes | 319 Views
Music retrieval. Conventional music retrieval systems Exact queries: ”Give me all songs from J.Lo’s latest album” What about ”Give me the music that I like”? New methods are needed: sophisticated similarity measures Increasing importance: MP3 players (10 3 songs)
E N D
Music retrieval • Conventional music retrieval systems • Exact queries: ”Give me all songs from J.Lo’s latest album” • What about ”Give me the music that I like”? New methods are needed: sophisticated similarity measures • Increasing importance: • MP3 players (103 songs) • Personal music collections (104 songs) • Music on demand • many songs, huge market value…
Proposal • Try a classifier method • Similarity measure enables matching of fuzzy data always returns results • Implement relevance feedback • User feedback Improves retrieval performance
Classifier systems • Genetic programming • Neural networks • Curve fitting algorithms • Vector quantizers
Tree structured Vector Quantization • Audio parameterization Feature space: MFCC coefficients • Quantization tree A supervised learning algorithm, TreeQ: • Attempts to partition feature space for maximum class separation
Features: MFCC coefficients waveform 100 Hamming windows/second DFT Log Mel IDFT MFCCs: A 13-dimensional vector per window 5 minutes song 30103 windows
Nearest neighbor Discrimination line in feature space • Problems: • Curse of dimensionality • Distribution assumptions • Complicated distributions
Each surface is added such that It cuts only one dimension (speed) the mutual information is maximized: Vector Quantization:Adding decision surfaces
Until further splits are not worthwile – according to certain stop conditions
Decision tree • Tree partitions features space • L regions (cells/leaves) • Based on class belonging of training data
Template generation • Generate templates for • Training data • Test data • Each MFCC vector is routed through the tree
Template generation • With a series of feature vectors, each vector will end up in one of the leaves. • This results in a histogram, or template, for each series of feature vectors.
Augmented similarity measure, e.g. DiffSim(X) = sim(X,A) –sim(X,C) Template comparison Corpus templates – one per training class A B n Query template Compute similarity X sim(X,A), sim(X,B), sim(X,C), …sim(X,n)
Result list Template comparison Corpus templates – one per training class A B n Query templates Compute similarity DiffSim(X) Sort
Preliminary experiments • Test subjects listened to 107 songs Rated them: good, fair, poor (class belonging Cg, Cf, Cp) • Training process: • For each user • Select randomly a subset (N songs) from each class • Construct a tree based on class belonging • Generate histogram templates for Cg, Cf, Cp • For each song X • Generate histogram template • Compute DiffSim(X) = sim(X,Cg) – sim(X,Cp) • Sort the list of songs according to DiffSim
Result list Relevance feedback classifier user
Implementation Adjust histogram profiles based on user feedback • For each user • Select the top M songs from the result list • Add the contents of the songs to the histogram profile based on the user rating (class belonging Cg, Cf, Cp) • For each song X • Generate histogram template • Compute DiffSim(X) = sim(X,Cg) – sim(X,Cp) • Sort the list of songs according to DiffSim