KNOWLEDGE TRANSFER FROM WEAKLY LABELED AUDIO USING CONVOLUTIONAL NEURAL NETWORK FOR SOUND EVENTS AND SCENES. pdf

Authors: Anurag Kumar, Maksim Khadkevich, Christian Fügen


Average Precision (AP)

average Precision

Figure compared AP of \(\mathcal{N}_S\) and \(\mathcal{N}_S^{slat}\). The Index of sound events 1 to 50 is shown below. The first column is Index in the above figure, second is the event id (magenta) as used in Audioset dataset and third (blue) is sound event name

1    /m/07qrkrw   Meow
2    /m/0dgbq   Civil defense siren
3    /m/0d31p   Vacuum cleaner
4    /m/01d380   Drill
5    /m/074ft   Song
6    /m/01w250   Whistling
7    /m/0dls3   Grunge
8    /m/07rwm0c   Clickety-clack
9    /m/01m2v   Computer keyboard
10    /m/03m5k   Harp
11    /m/03fwl   Goat
12    /m/07s8j8t   Roll
13    /m/018w8   Basketball bounce
14    /m/07q0h5t   Bleat
15    /m/03q5t   Harpsichord
16    /m/06j64v   Middle Eastern music
17    /m/07qn5dc   Crowing- cock-a-doodle-doo
18    /m/0316dw   Typing
19    /m/06wzb   Steam
20    /m/01wy6   Clarinet
21    /m/0140xf   Christmas music
22    /m/01sm1g   Wood block
23    /m/03qjg   Harmonica
24    /m/02w4v   Folk music
25    /m/07c52   Television
26    /m/01jt3m   Toilet flush
27    /m/028ght   Applause
28    /m/01h8n0   Conversation
29    /m/014zdl   Explosion
30    /m/02fsn   Double bass
31    /m/0284vy3   Train horn
32    /m/05lls   Opera
33    /m/07qsvvw   Burst- pop
34    /m/05w3f   Psychedelic rock
35    /m/01d3sd   Snoring
36    /m/07r_k2n   Yip
37    /m/0326g   Flamenco
38    /m/04rzd   Mandolin
39    /t/dd00002   Baby cry- infant cry
40    /m/02jz0l   Water tap- faucet
41    /m/01jwx6   Vibration
42    /m/018j2   Banjo
43    /m/02cz_7   Beatboxing
44    /m/07st89h   Cluck
45    /m/02dgv   Door
46    /m/0dq0md   Music of Bollywood
47    /m/0ln16   Salsa music
48    /m/07qdb04   Quack
49    /m/07rjwbb   Hiss
50    /m/09ddx   Duck