Where should a conference paper extend?

Follow the extended contribution: vision depth → TPAMI; learning-systems novelty → TNNLS.

Compare

TNNLS vs TPAMI · scope and template differences

Both are top transactions—with different readers and template families. The wrong shell hurts more than a slightly wrong page guess.

Facet	TNNLS	TPAMI
Best-fit contribution	Neural / learning-systems methods	Vision / pattern-analysis long form
Template	Standard IEEE journal (not compsoc)	IEEE Computer Society template
Length (typical)	Mandatory overlength often after ~10 pages	Regular often ~12 pages incl. refs

Decide learning-systems vs vision/pattern-analysis first, download the matching official template, then worry about fees and OA.

Where should a conference paper extend?: Follow the extended contribution: vision depth → TPAMI; learning-systems novelty → TNNLS.