Closed niconoe closed 3 years ago
In GitLab by @kbubnicki on Aug 27, 2020, 00:22
I can not agree here. The term event
is quite common in camera trapping literature, see e.g. Meek et al. 2014:
The term ‘‘event’’ typically refers to the presence of an animal, on a single occasion in a
location, as detected by a sensor and documented by a camera. Two factors can determine
the precise definition of a camera trap event and should be clarified by the investigators:
independence interval between successive images or sequences of consecutive images (i.e.
video or pseudo video) initiated by a trigger. Although the optimal independence interval
has never been empirically estimated from camera traps, it typically ranges from 10 to
60 min (although 1–5 min has been used for small mammals), and the researcher should
report on the criteria used to set the interval.
So if set to 30 seconds, then images A (20:30:00), B (20:30:15), C (20:32:00) would be considered as 2 sequences: A+B and C?
-> Yes!
But I find the term sequence_cutoff
much more confusing.
In GitLab by @peterdesmet on Aug 27, 2020, 10:26
I agree that the term event
is quite common for biological data (also used in Darwin Core).
Question: do we consider event
and sequence
to be different names for the same concept?
If so, then I think we should use the same term consistently:
event_interval
event_id
, event_start
, event_end
.If not, can you clarify the difference between event and sequence?
In GitLab by @kbubnicki on Aug 28, 2020, 12:09
Good question. I would say that event
is an actual observation i.e. a data point whereas sequence
is more a technical/grouping term for a consecutive set of images recorded at one time. But what actually matters in the end is event
. The question is if we need both terms or not (if so this can be confusing). I would opt for keeping only event
.
In TRAPPER we use only event
which is actually named sequence
:)
In GitLab by @peterdesmet on Aug 28, 2020, 12:35
@yliefting, @milotictanja could you agree with the terminology that we organize media (and related observations) in event
s rather than sequence
s?
In GitLab by @yliefting on Aug 28, 2020, 13:06
The main reason we use the term sequence in Agouti is that it, like @kbubnicki already said, is just a simple grouping of images based on the difference in timestamps. Event implies that it is a data point with meaning. Theoretically a sequence can have multiple events, described in several observations. Since there are large differences between project protocols we chose to stick with sequences and leave any interpretation to the users.
That said, I have no objections to using the term event in the standard. We can simply map Agouti's sequences as events.
One remark: in Agouti videos are always mapped as one video equals one sequence The analogy here is that a video is a sequence of frames, and photos taken in bursts are similar.
In GitLab by @kbubnicki on Aug 28, 2020, 13:20
Ok, thanks @yliefting! This is all so confusing as theoretically one event can also have multiple sequences :) One question: is the time interval to build sequences (sequence_cutoff
?) user-defined in Agouti?
In GitLab by @yliefting on Aug 28, 2020, 14:14
Haha, you're right. Good thing we are working on this standard. ;)
The sequence cutoff in Agouti is defined by the project owner at the project level and can vary. The default is 120 seconds, but some projects deviate because they have specific requirements.
In GitLab by @peterdesmet on Aug 28, 2020, 14:17
Do we have a consensus of generalizing the term sequence
to event
? Thumbs up if you agree, otherwise please continue discussion. :smiley:
In GitLab by @peterdesmet on Aug 31, 2020, 17:19
mentioned in commit b172b07589fa560eb9245be3007fea3badb3c6db
In GitLab by @peterdesmet on Aug 31, 2020, 17:26
mentioned in merge request !4
In GitLab by @kbubnicki on Sep 2, 2020, 14:20
I think that this one we can also close now.
In GitLab by @peterdesmet on Sep 2, 2020, 15:46
changed title from Clarify definition of event_interval to Clarify definition of eventinterval{+, use event throughout+}
In GitLab by @peterdesmet on Sep 2, 2020, 15:47
assigned to @peterdesmet and unassigned @kbubnicki
In GitLab by @peterdesmet on Sep 2, 2020, 15:48
Consensus: all sequence_
proposed terms should be event_
. The definition of event_interval
could be clarified a bit further, will take a stab at this.
In GitLab by @peterdesmet on Sep 2, 2020, 17:00
My proposal for event_interval
:
Maximum number of seconds between timestamps of successive multimedia files to be considered part of a single independent event/occasion and be assigned the same
event_id
.
Note: I still find the word event_interval
a bit confusing, as it seems to be the interval between events. Maybe event_media_interval
?
I would also update event_id
to reflect :
A unique identifier of an event within a project. Events contain one or more multimedia files (e.g. a single video or a sequence of successive images) and are defined by
event_media_interval
in the package metadata.
Note: successive vs consecutive (1,2,4 is successive, not consecutive, while 1,2,3 is both)
Updated from:
A unique identifier of an event within a project. Events are what project researchers or software have defined as independent occasions, containing one or more multimedia files (e.g. a single video or a sequence of consecutive images).
In GitLab by @kbubnicki on Sep 2, 2020, 17:13
Already cited Meek et al. 2014 used a term independence interval
. I have no idea now, so maybe again lets just keep event_interval
and ask the community afterwards?
In GitLab by @peterdesmet on Sep 2, 2020, 20:00
Ok.
In GitLab by @kbubnicki on Sep 3, 2020, 00:41
mentioned in commit 1cf8d960d276ce21e6b3f2f75e918b241f04545d
In GitLab by @kbubnicki on Sep 3, 2020, 00:46
mentioned in merge request !9
In GitLab by @peterdesmet on Sep 4, 2020, 13:31
Implemented in !9.
In GitLab by @peterdesmet on Aug 7, 2020, 10:05
The package property
event_interval
is defined as:sequence_cutoff
.sequence_cutoff
? I find that clearer to understand.