diff mbox series

[FFmpeg-devel,2/2] libavfilter/signature_lookup: fix jaccard distance

Message ID 20240505223040.521838-2-gerion.entrup@flump.de
State New
Headers show
Series [FFmpeg-devel,1/2] libavfilter/signature_lookup: fix possible division by zero | expand

Checks

Context Check Description
andriy/make_x86 success Make finished
andriy/make_fate_x86 success Make fate finished

Commit Message

Gerion Entrup May 5, 2024, 10:30 p.m. UTC 
Actually, the jaccard distance is defined as D = 1 - intersect / union.
Additionally, the distance value is compared against a constant that
must be between 0 and 1, which is not the case here. Both facts together
has led to the fact, that the function always returned a matching course
signature. To leave the constant intact and to avoid floating point
computation, this commit multiplies with 1 << 16 making the constant
effectively 9000 / (1<<16) =~ 0.14.

Reported-by: Sachin Tilloo <sachin.tilloo@gmail.com>
Reviewed-by: Sachin Tilloo <sachin.tilloo@gmail.com>
Tested-by: Sachin Tilloo <sachin.tilloo@gmail.com>
---
 libavfilter/signature_lookup.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff mbox series

Patch

diff --git a/libavfilter/signature_lookup.c b/libavfilter/signature_lookup.c
index b39a3e225b..b90b63f3f2 100644
--- a/libavfilter/signature_lookup.c
+++ b/libavfilter/signature_lookup.c
@@ -127,9 +127,10 @@  static int get_jaccarddist(SignatureContext *sc, CoarseSignature *first, CoarseS
 {
     int jaccarddist, i, composdist = 0, cwthcount = 0;
     for (i = 0; i < 5; i++) {
-        if ((jaccarddist = intersection_word(first->data[i], second->data[i])) > 0) {
+        if ((jaccarddist = (1 << 16) * intersection_word(first->data[i], second->data[i])) > 0) {
             jaccarddist /= FFMAX(union_word(first->data[i], second->data[i]), 1);
         }
+        jaccarddist = (1 << 16) - jaccarddist;
         if (jaccarddist >= sc->thworddist) {
             if (++cwthcount > 2) {
                 /* more than half (5/2) of distances are too wide */